New results

pythia-14m-seed0/step0/EleutherAI__pythia-14m/results_2024-08-15T12-48-54.868316.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 3507781.8105202965,
+      "perplexity_stderr,none": 338907.2426000304,
+      "acc,none": 0.0,
+      "acc_stderr,none": 0.0,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step0,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step0",
+    "model_sha": "bc249c50b4a381f99e9296f08a69b8562a31fde4",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723751264.2585666,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1133831.651542538,
+  "end_time": 1133915.107703607,
+  "total_evaluation_time_seconds": "83.45616106898524"
+}

pythia-14m-seed0/step1/EleutherAI__pythia-14m/results_2024-08-15T12-50-21.093189.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 3507781.8105202965,
+      "perplexity_stderr,none": 338907.2426000304,
+      "acc,none": 0.0,
+      "acc_stderr,none": 0.0,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step1,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step1",
+    "model_sha": "71caa8178c55d927b2234a30bbabac6bd072cae8",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723751353.851652,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1133922.894150664,
+  "end_time": 1134001.333325669,
+  "total_evaluation_time_seconds": "78.43917500483803"
+}

pythia-14m-seed0/step1000/EleutherAI__pythia-14m/results_2024-08-15T13-04-58.314501.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 195988.90526637834,
+      "perplexity_stderr,none": 12193.7268807034,
+      "acc,none": 0.0,
+      "acc_stderr,none": 0.0,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step1000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step1000",
+    "model_sha": "67ed4908dc8d91741d246c4d9fbfa03e78c8505d",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723752224.1879683,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1134793.204587141,
+  "end_time": 1134878.555076991,
+  "total_evaluation_time_seconds": "85.35048984992318"
+}

pythia-14m-seed0/step10000/EleutherAI__pythia-14m/results_2024-08-15T13-18-05.667000.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1511.251720455259,
+      "perplexity_stderr,none": 70.29803394374686,
+      "acc,none": 0.1003299049097613,
+      "acc_stderr,none": 0.004185707878554855,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step10000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step10000",
+    "model_sha": "50c427ff77acdb2553ee434ffabe182f5be3c1e2",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723753014.309483,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1135582.228500084,
+  "end_time": 1135665.902803599,
+  "total_evaluation_time_seconds": "83.67430351488292"
+}

pythia-14m-seed0/step100000/EleutherAI__pythia-14m/results_2024-08-15T15-27-54.136872.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2480.572303985594,
+      "perplexity_stderr,none": 110.31519283833292,
+      "acc,none": 0.05142635358043858,
+      "acc_stderr,none": 0.0030770918924100804,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step100000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step100000",
+    "model_sha": "c5ab9636d47fcd482787a66ae2ef8a71002648b9",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723760806.4035344,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143373.76838171,
+  "end_time": 1143454.366109349,
+  "total_evaluation_time_seconds": "80.59772763913497"
+}

pythia-14m-seed0/step101000/EleutherAI__pythia-14m/results_2024-08-15T15-29-20.379855.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1936.7673989556679,
+      "perplexity_stderr,none": 89.14269698503178,
+      "acc,none": 0.07354938870560838,
+      "acc_stderr,none": 0.0036367440978540603,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step101000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step101000",
+    "model_sha": "2a340986858902da1c8563a6c273066aaccc8750",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723760893.3566766,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143461.457184083,
+  "end_time": 1143540.619783163,
+  "total_evaluation_time_seconds": "79.16259908000939"
+}

pythia-14m-seed0/step102000/EleutherAI__pythia-14m/results_2024-08-15T15-30-45.504373.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2332.1576555643323,
+      "perplexity_stderr,none": 103.2466067379606,
+      "acc,none": 0.055695711236173104,
+      "acc_stderr,none": 0.0031950593588073433,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step102000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step102000",
+    "model_sha": "e1ef05f9ceddbf4c8b5eda7b6e1eb66da607dc76",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723760978.5723045,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143547.70613283,
+  "end_time": 1143625.744282615,
+  "total_evaluation_time_seconds": "78.0381497850176"
+}

pythia-14m-seed0/step103000/EleutherAI__pythia-14m/results_2024-08-15T15-32-11.742196.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2799.559936979982,
+      "perplexity_stderr,none": 123.34505192306801,
+      "acc,none": 0.05491946438967592,
+      "acc_stderr,none": 0.0031740197788982893,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step103000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step103000",
+    "model_sha": "2f4b2967a8d878296de9228ba494cd18a59fe35c",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761064.4161794,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143633.446766386,
+  "end_time": 1143711.9823575,
+  "total_evaluation_time_seconds": "78.53559111384675"
+}

pythia-14m-seed0/step104000/EleutherAI__pythia-14m/results_2024-08-15T15-33-38.569825.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2003.2493640780494,
+      "perplexity_stderr,none": 89.98672601843188,
+      "acc,none": 0.06617504366388512,
+      "acc_stderr,none": 0.003463314484220094,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step104000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step104000",
+    "model_sha": "678191556681a68fbcd641ab2bd018f3634fd545",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761150.9524848,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143719.967749508,
+  "end_time": 1143798.810041786,
+  "total_evaluation_time_seconds": "78.8422922778409"
+}

pythia-14m-seed0/step105000/EleutherAI__pythia-14m/results_2024-08-15T15-35-05.211421.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2204.366599640772,
+      "perplexity_stderr,none": 101.36401689274162,
+      "acc,none": 0.06966815447312245,
+      "acc_stderr,none": 0.003546893672151897,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step105000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step105000",
+    "model_sha": "52a764e3d099f01bcf6079282458009fba8e9815",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761237.5477817,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143806.663444182,
+  "end_time": 1143885.451216583,
+  "total_evaluation_time_seconds": "78.78777240100317"
+}

pythia-14m-seed0/step106000/EleutherAI__pythia-14m/results_2024-08-15T15-36-32.465444.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2260.64934437262,
+      "perplexity_stderr,none": 101.92424646136189,
+      "acc,none": 0.06986221618474675,
+      "acc_stderr,none": 0.003551459735491157,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step106000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step106000",
+    "model_sha": "fb05d722aeeb0944d2f041429d68cd7a0f0aefb5",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761324.5612192,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143892.700193422,
+  "end_time": 1143972.705910608,
+  "total_evaluation_time_seconds": "80.00571718602441"
+}

pythia-14m-seed0/step107000/EleutherAI__pythia-14m/results_2024-08-15T15-38-00.033086.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2030.1912500934511,
+      "perplexity_stderr,none": 89.98267591591406,
+      "acc,none": 0.05918882204541044,
+      "acc_stderr,none": 0.003287631671733364,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step107000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step107000",
+    "model_sha": "e51d508ae0521ac7f4d1f9879ae7a45133c4c71f",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761412.6764872,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1143979.738471963,
+  "end_time": 1144060.273198181,
+  "total_evaluation_time_seconds": "80.5347262180876"
+}

pythia-14m-seed0/step108000/EleutherAI__pythia-14m/results_2024-08-15T15-39-28.787439.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1942.191674587457,
+      "perplexity_stderr,none": 88.17936208188293,
+      "acc,none": 0.07219095672423831,
+      "acc_stderr,none": 0.003605643396603442,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step108000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step108000",
+    "model_sha": "e6e41d34d4cea7f1b7b7a3c789f10c07e9ea998a",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761500.0742729,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144067.406098665,
+  "end_time": 1144149.027880051,
+  "total_evaluation_time_seconds": "81.62178138596937"
+}

pythia-14m-seed0/step109000/EleutherAI__pythia-14m/results_2024-08-15T15-40-55.857604.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1856.8399370102884,
+      "perplexity_stderr,none": 85.6448562808026,
+      "acc,none": 0.07413157384048127,
+      "acc_stderr,none": 0.0036499618074347898,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step109000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step109000",
+    "model_sha": "1406e20c840393345a50b6b8eaa261e2a8afa13f",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761588.1894023,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144156.112510001,
+  "end_time": 1144236.097111593,
+  "total_evaluation_time_seconds": "79.98460159194656"
+}

pythia-14m-seed0/step11000/EleutherAI__pythia-14m/results_2024-08-15T13-19-30.603325.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1336.2520161255652,
+      "perplexity_stderr,none": 60.98331616528402,
+      "acc,none": 0.10285270716087716,
+      "acc_stderr,none": 0.004232059957170525,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step11000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step11000",
+    "model_sha": "23b9f9586a7ce1aa3b85566361c5471f05e36f30",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723753103.9197896,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1135673.004151162,
+  "end_time": 1135750.843636572,
+  "total_evaluation_time_seconds": "77.83948540990241"
+}

pythia-14m-seed0/step110000/EleutherAI__pythia-14m/results_2024-08-15T15-42-21.675993.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1655.1447995081642,
+      "perplexity_stderr,none": 76.78861282594242,
+      "acc,none": 0.08092373374733165,
+      "acc_stderr,none": 0.0037994943646444323,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step110000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step110000",
+    "model_sha": "4f102eb31047258ae88f77a33ad7a1d22107ad7f",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761674.655817,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144243.183542134,
+  "end_time": 1144321.915977644,
+  "total_evaluation_time_seconds": "78.73243551002815"
+}

pythia-14m-seed0/step111000/EleutherAI__pythia-14m/results_2024-08-15T15-43-46.617900.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1577.7555976392246,
+      "perplexity_stderr,none": 73.319450716533,
+      "acc,none": 0.08286435086357462,
+      "acc_stderr,none": 0.0038407207037192497,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step111000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step111000",
+    "model_sha": "99137989374bfe1ecb1a5e64b3d80ce774eec113",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761759.958622,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144328.95044279,
+  "end_time": 1144406.857867798,
+  "total_evaluation_time_seconds": "77.90742500801571"
+}

pythia-14m-seed0/step112000/EleutherAI__pythia-14m/results_2024-08-15T15-45-12.514667.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1759.9239677979742,
+      "perplexity_stderr,none": 81.70703449195138,
+      "acc,none": 0.07490782068697846,
+      "acc_stderr,none": 0.0036674834158285856,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step112000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step112000",
+    "model_sha": "9774065c9f5a45ad7788ad185541011cbb0862d9",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761845.0600786,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144414.160817905,
+  "end_time": 1144492.75489787,
+  "total_evaluation_time_seconds": "78.59407996479422"
+}

pythia-14m-seed0/step113000/EleutherAI__pythia-14m/results_2024-08-15T15-46-38.770492.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1588.236297320137,
+      "perplexity_stderr,none": 72.93322208929094,
+      "acc,none": 0.07490782068697846,
+      "acc_stderr,none": 0.0036674834158285856,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step113000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step113000",
+    "model_sha": "550e364271f369edb9d4a144425830fdedee75cd",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723761930.9352212,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144499.950961815,
+  "end_time": 1144579.010207957,
+  "total_evaluation_time_seconds": "79.05924614192918"
+}

pythia-14m-seed0/step114000/EleutherAI__pythia-14m/results_2024-08-15T15-48-05.260016.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1702.065327436129,
+      "perplexity_stderr,none": 78.56178715398188,
+      "acc,none": 0.07568406753347565,
+      "acc_stderr,none": 0.0036848899715327707,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step114000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step114000",
+    "model_sha": "9c800cd9948d544aff8957bed479581938a7d59c",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762017.8288126,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144586.610589841,
+  "end_time": 1144665.499988281,
+  "total_evaluation_time_seconds": "78.88939844002016"
+}

pythia-14m-seed0/step115000/EleutherAI__pythia-14m/results_2024-08-15T15-49-32.634482.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1653.5844188453527,
+      "perplexity_stderr,none": 76.53860729699839,
+      "acc,none": 0.07471375897535416,
+      "acc_stderr,none": 0.003663113864225891,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step115000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step115000",
+    "model_sha": "84565561177046e994cc5d236a3f526cb0a1494e",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762104.9888885,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144673.793967198,
+  "end_time": 1144752.874586123,
+  "total_evaluation_time_seconds": "79.08061892492697"
+}

pythia-14m-seed0/step116000/EleutherAI__pythia-14m/results_2024-08-15T15-51-01.124488.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1670.6856831411806,
+      "perplexity_stderr,none": 76.90491802120467,
+      "acc,none": 0.07432563555210557,
+      "acc_stderr,none": 0.0036543530860415506,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step116000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step116000",
+    "model_sha": "90fd8b63997a429c96f4d13a9d5290cb9621af6e",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762193.411083,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144760.733506932,
+  "end_time": 1144841.363911434,
+  "total_evaluation_time_seconds": "80.6304045021534"
+}

pythia-14m-seed0/step117000/EleutherAI__pythia-14m/results_2024-08-15T15-52-30.240629.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1754.3725242092196,
+      "perplexity_stderr,none": 81.01434898802123,
+      "acc,none": 0.08072967203570736,
+      "acc_stderr,none": 0.003795336508106496,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step117000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step117000",
+    "model_sha": "b205c539eccb75ec24cffe6c918b8eb5419e9e1f",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762281.6944668,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144849.179324036,
+  "end_time": 1144930.480355115,
+  "total_evaluation_time_seconds": "81.30103107891046"
+}

pythia-14m-seed0/step118000/EleutherAI__pythia-14m/results_2024-08-15T15-53-56.953504.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1758.961983552553,
+      "perplexity_stderr,none": 80.65986006671504,
+      "acc,none": 0.07587812924509994,
+      "acc_stderr,none": 0.0036892238245559935,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step118000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step118000",
+    "model_sha": "0cd8490766c1ec80c1f9aa970de646bd573192d1",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762369.5567913,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1144937.75132631,
+  "end_time": 1145017.193742402,
+  "total_evaluation_time_seconds": "79.44241609191522"
+}

pythia-14m-seed0/step119000/EleutherAI__pythia-14m/results_2024-08-15T15-55-23.128925.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1654.0146622462084,
+      "perplexity_stderr,none": 76.2436881269689,
+      "acc,none": 0.07665437609159713,
+      "acc_stderr,none": 0.0037064888417067503,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step119000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step119000",
+    "model_sha": "06f016ef45fd4df8e4118bec1aaaa17ce539d8eb",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762455.409844,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145024.27587102,
+  "end_time": 1145103.369137769,
+  "total_evaluation_time_seconds": "79.09326674905606"
+}

pythia-14m-seed0/step12000/EleutherAI__pythia-14m/results_2024-08-15T13-20-56.302395.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1345.9520903737475,
+      "perplexity_stderr,none": 62.175751078363405,
+      "acc,none": 0.09179118959829226,
+      "acc_stderr,none": 0.004022586828929574,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step12000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step12000",
+    "model_sha": "677c73e05b0353916ad35c85c4f4609dce0348af",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723753189.099923,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2200.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1135757.985463659,
+  "end_time": 1135836.540199971,
+  "total_evaluation_time_seconds": "78.55473631201312"
+}

pythia-14m-seed0/step120000/EleutherAI__pythia-14m/results_2024-08-15T15-56-52.216443.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1885.6556307949313,
+      "perplexity_stderr,none": 84.00187419647823,
+      "acc,none": 0.06462254997089074,
+      "acc_stderr,none": 0.003425291731867342,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step120000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step120000",
+    "model_sha": "379be14e11d7e3ef68044bbbcc8890e192c013b9",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762541.51547,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145110.479581879,
+  "end_time": 1145192.456666261,
+  "total_evaluation_time_seconds": "81.97708438197151"
+}

pythia-14m-seed0/step121000/EleutherAI__pythia-14m/results_2024-08-15T15-58-17.456909.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1760.6364312354708,
+      "perplexity_stderr,none": 81.10863398516484,
+      "acc,none": 0.07781874636134291,
+      "acc_stderr,none": 0.0037321778637123253,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step121000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step121000",
+    "model_sha": "b4462c932e24f3271ec4e0819a6af0af31105a9a",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762630.560148,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145199.569018539,
+  "end_time": 1145277.697085722,
+  "total_evaluation_time_seconds": "78.12806718284264"
+}

pythia-14m-seed0/step122000/EleutherAI__pythia-14m/results_2024-08-15T15-59-44.315496.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1614.3906537485614,
+      "perplexity_stderr,none": 75.54730571478007,
+      "acc,none": 0.0836405977100718,
+      "acc_stderr,none": 0.0038570347957365973,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step122000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step122000",
+    "model_sha": "d05207ca4e2e9503facd5d5f03130fec1accc174",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762715.8275197,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145284.959185133,
+  "end_time": 1145364.550398592,
+  "total_evaluation_time_seconds": "79.59121345891617"
+}

pythia-14m-seed0/step123000/EleutherAI__pythia-14m/results_2024-08-15T16-01-10.335351.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2038.1899485513527,
+      "perplexity_stderr,none": 91.65050326709839,
+      "acc,none": 0.066757228798758,
+      "acc_stderr,none": 0.0034774311506045566,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step123000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step123000",
+    "model_sha": "bf4ec8affb1219b69a78d3ea02f4300e65c2627e",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762802.8322046,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145371.838932145,
+  "end_time": 1145450.57434153,
+  "total_evaluation_time_seconds": "78.73540938505903"
+}

pythia-14m-seed0/step124000/EleutherAI__pythia-14m/results_2024-08-15T16-02-35.938574.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1899.7256494472654,
+      "perplexity_stderr,none": 85.27743433714627,
+      "acc,none": 0.0671453522220066,
+      "acc_stderr,none": 0.00348680002818163,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step124000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step124000",
+    "model_sha": "e6fb05f7d7a0437fe0bed7c289eeec6034f3ec93",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762888.69861,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145457.648579555,
+  "end_time": 1145536.161768229,
+  "total_evaluation_time_seconds": "78.51318867388181"
+}

pythia-14m-seed0/step125000/EleutherAI__pythia-14m/results_2024-08-15T16-04-02.754305.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1716.8749780802998,
+      "perplexity_stderr,none": 79.70990558710086,
+      "acc,none": 0.07607219095672424,
+      "acc_stderr,none": 0.0036935506133658606,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step125000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step125000",
+    "model_sha": "460b51a3b8394f0bbcc5f839d989f2e19a262a9f",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723762975.4865017,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145543.265744231,
+  "end_time": 1145622.994148452,
+  "total_evaluation_time_seconds": "79.72840422112495"
+}

pythia-14m-seed0/step126000/EleutherAI__pythia-14m/results_2024-08-15T16-05-29.292044.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1648.377668467864,
+      "perplexity_stderr,none": 76.19708183106972,
+      "acc,none": 0.07704249951484572,
+      "acc_stderr,none": 0.003715079456679115,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step126000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step126000",
+    "model_sha": "7bdc419ff72691c556674155c1a83bc6e88cd742",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763062.0582552,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145630.007922808,
+  "end_time": 1145709.532083368,
+  "total_evaluation_time_seconds": "79.52416055998765"
+}

pythia-14m-seed0/step127000/EleutherAI__pythia-14m/results_2024-08-15T16-06-56.433830.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1746.9065161719782,
+      "perplexity_stderr,none": 80.67923580716102,
+      "acc,none": 0.07898311663108869,
+      "acc_stderr,none": 0.003757621238955833,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step127000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step127000",
+    "model_sha": "0ebce7cc0fad8445578b776dae9f0c2dd9017d50",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763149.0984063,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145716.698455785,
+  "end_time": 1145796.673910495,
+  "total_evaluation_time_seconds": "79.97545470995829"
+}

pythia-14m-seed0/step128/EleutherAI__pythia-14m/results_2024-08-15T13-00-34.735877.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 2154450.549794474,
+      "perplexity_stderr,none": 190529.88059195384,
+      "acc,none": 0.0,
+      "acc_stderr,none": 0.0,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step128,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step128",
+    "model_sha": "712cec71d5646cb88f76ee92029c12132ccd754c",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723751967.8752835,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1134536.513561532,
+  "end_time": 1134614.975972693,
+  "total_evaluation_time_seconds": "78.46241116104648"
+}

pythia-14m-seed0/step128000/EleutherAI__pythia-14m/results_2024-08-15T16-08-22.902858.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1695.180554031583,
+      "perplexity_stderr,none": 77.5396989490735,
+      "acc,none": 0.06908596933824956,
+      "acc_stderr,none": 0.003533147663272781,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step128000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step128000",
+    "model_sha": "698a9b44b5707c15193cb9462f02c1952bf7ef29",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763234.9688888,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145803.817017457,
+  "end_time": 1145883.143051203,
+  "total_evaluation_time_seconds": "79.32603374612518"
+}

pythia-14m-seed0/step129000/EleutherAI__pythia-14m/results_2024-08-15T16-09-48.457182.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1748.2060562452823,
+      "perplexity_stderr,none": 80.70472934165429,
+      "acc,none": 0.06908596933824956,
+      "acc_stderr,none": 0.003533147663272781,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step129000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step129000",
+    "model_sha": "9e3f931ba445a319b202efae7ca3b6624b085ae4",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763321.257125,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145890.282894991,
+  "end_time": 1145968.695761718,
+  "total_evaluation_time_seconds": "78.41286672698334"
+}

pythia-14m-seed0/step13000/EleutherAI__pythia-14m/results_2024-08-15T13-22-24.295538.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1245.8883648802528,
+      "perplexity_stderr,none": 57.69527819537371,
+      "acc,none": 0.10576363283524161,
+      "acc_stderr,none": 0.004284561852149411,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step13000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step13000",
+    "model_sha": "51725d80c402ddd0b1ef23cb322f0fff2b7ab057",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723753276.8532705,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1135845.439155096,
+  "end_time": 1135924.535229917,
+  "total_evaluation_time_seconds": "79.0960748209618"
+}

pythia-14m-seed0/step130000/EleutherAI__pythia-14m/results_2024-08-15T16-11-14.614730.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1615.9022166310167,
+      "perplexity_stderr,none": 74.77044537396877,
+      "acc,none": 0.08111779545895595,
+      "acc_stderr,none": 0.0038036457543544435,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step130000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step130000",
+    "model_sha": "c5a452c0e21b2beb5583459f763449ed56caa5c2",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763407.04308,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1145975.991028769,
+  "end_time": 1146054.855413207,
+  "total_evaluation_time_seconds": "78.86438443814404"
+}

pythia-14m-seed0/step131000/EleutherAI__pythia-14m/results_2024-08-15T16-12-40.692551.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1429.6528872004685,
+      "perplexity_stderr,none": 66.43871183549871,
+      "acc,none": 0.08888026392392781,
+      "acc_stderr,none": 0.003964628217610132,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step131000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step131000",
+    "model_sha": "4facff6a23be64c8b725719006ddd396b322144f",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763493.1769783,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146062.15420514,
+  "end_time": 1146140.932159504,
+  "total_evaluation_time_seconds": "78.77795436396264"
+}

pythia-14m-seed0/step132000/EleutherAI__pythia-14m/results_2024-08-15T16-14-06.187317.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1744.4924353855185,
+      "perplexity_stderr,none": 81.16053763865396,
+      "acc,none": 0.07180283330098972,
+      "acc_stderr,none": 0.0035966897962539787,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step132000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step132000",
+    "model_sha": "d32e674e5bcaaabda91936dc3c8d6ebc2873fa18",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763579.0232463,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146148.026271967,
+  "end_time": 1146226.427460756,
+  "total_evaluation_time_seconds": "78.40118878893554"
+}

pythia-14m-seed0/step133000/EleutherAI__pythia-14m/results_2024-08-15T16-15-33.241417.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1654.4108715342993,
+      "perplexity_stderr,none": 77.65858495677924,
+      "acc,none": 0.08228216572870173,
+      "acc_stderr,none": 0.0038284194650137684,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step133000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step133000",
+    "model_sha": "c0fe6645a6d36cae5975f2d5ba418a058ab1023e",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763665.9182217,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146233.538208775,
+  "end_time": 1146313.481529114,
+  "total_evaluation_time_seconds": "79.94332033907995"
+}

pythia-14m-seed0/step134000/EleutherAI__pythia-14m/results_2024-08-15T16-17-00.743542.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1714.602030249706,
+      "perplexity_stderr,none": 78.98081491097976,
+      "acc,none": 0.07762468464971861,
+      "acc_stderr,none": 0.003727913555390672,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step134000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step134000",
+    "model_sha": "6aacc13c38d2aa88207071a8868e41c6e22cc82b",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763753.4132254,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146320.668391112,
+  "end_time": 1146400.983705253,
+  "total_evaluation_time_seconds": "80.31531414110214"
+}

pythia-14m-seed0/step135000/EleutherAI__pythia-14m/results_2024-08-15T16-18-28.514020.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1645.9272622179124,
+      "perplexity_stderr,none": 76.43349328751616,
+      "acc,none": 0.07743062293809432,
+      "acc_stderr,none": 0.0037236424005269012,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step135000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step135000",
+    "model_sha": "95509b738e052a84f0d6741472fbec47268ada6d",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763840.8905025,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146408.240149021,
+  "end_time": 1146488.754164905,
+  "total_evaluation_time_seconds": "80.51401588413864"
+}

pythia-14m-seed0/step136000/EleutherAI__pythia-14m/results_2024-08-15T16-19-54.399480.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1960.7084303102972,
+      "perplexity_stderr,none": 88.24537878239123,
+      "acc,none": 0.06442848825926645,
+      "acc_stderr,none": 0.003420499552612167,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step136000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step136000",
+    "model_sha": "2dd2b497d1fa1285e9288538bf03af70ea88c3f1",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723763927.2161357,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146495.818538224,
+  "end_time": 1146574.639468531,
+  "total_evaluation_time_seconds": "78.82093030679971"
+}

pythia-14m-seed0/step137000/EleutherAI__pythia-14m/results_2024-08-15T16-21-20.940162.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1592.062887278126,
+      "perplexity_stderr,none": 73.0489453312968,
+      "acc,none": 0.07762468464971861,
+      "acc_stderr,none": 0.003727913555390672,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step137000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step137000",
+    "model_sha": "f9cd17305d5cfb4f3776c1ea96d9b0f00fe32ede",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723764012.7023847,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146581.770022683,
+  "end_time": 1146661.180396564,
+  "total_evaluation_time_seconds": "79.41037388099357"
+}

pythia-14m-seed0/step138000/EleutherAI__pythia-14m/results_2024-08-15T16-22-46.483201.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1616.2517548162928,
+      "perplexity_stderr,none": 74.09170027590162,
+      "acc,none": 0.07490782068697846,
+      "acc_stderr,none": 0.0036674834158285856,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step138000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step138000",
+    "model_sha": "720f854dbe30d974b7b495c0b076e452aab854b9",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723764099.134047,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146668.278176658,
+  "end_time": 1146746.723473584,
+  "total_evaluation_time_seconds": "78.44529692595825"
+}

pythia-14m-seed0/step139000/EleutherAI__pythia-14m/results_2024-08-15T16-24-12.061095.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1698.8500597374266,
+      "perplexity_stderr,none": 78.62022294671738,
+      "acc,none": 0.07549000582185135,
+      "acc_stderr,none": 0.00368054902934182,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step139000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step139000",
+    "model_sha": "7dde889802ebc8e9b3498bcc7d02ce199d9d4d76",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723764184.9015696,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146753.933885823,
+  "end_time": 1146832.300943617,
+  "total_evaluation_time_seconds": "78.36705779400654"
+}

pythia-14m-seed0/step14000/EleutherAI__pythia-14m/results_2024-08-15T13-23-49.471021.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1334.3251049658206,
+      "perplexity_stderr,none": 62.74215047810956,
+      "acc,none": 0.10304676887250146,
+      "acc_stderr,none": 0.004235592410767947,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step14000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step14000",
+    "model_sha": "d1cd73f742163cf3d733776a53422349d945caa8",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723753362.6691036,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               2500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1135931.618820009,
+  "end_time": 1136009.711027219,
+  "total_evaluation_time_seconds": "78.09220720990561"
+}

pythia-14m-seed0/step140000/EleutherAI__pythia-14m/results_2024-08-15T16-25-37.975454.json

@@ -0,0 +1,96 @@
+{
+  "results": {
+    "lambada_openai": {
+      "perplexity,none": 1635.2112969586547,
+      "perplexity_stderr,none": 75.2598585640202,
+      "acc,none": 0.07490782068697846,
+      "acc_stderr,none": 0.0036674834158285856,
+      "alias": "lambada_openai"
+    }
+  },
+  "group_subtasks": {
+    "lambada_openai": []
+  },
+  "configs": {
+    "lambada_openai": {
+      "task": "lambada_openai",
+      "group": [
+        "lambada"
+      ],
+      "dataset_path": "EleutherAI/lambada_openai",
+      "dataset_name": "default",
+      "dataset_kwargs": {
+        "trust_remote_code": true
+      },
+      "test_split": "test",
+      "doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
+      "doc_to_target": "{{' '+text.split(' ')[-1]}}",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "perplexity",
+          "aggregation": "perplexity",
+          "higher_is_better": false
+        },
+        {
+          "metric": "acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "loglikelihood",
+      "repeats": 1,
+      "should_decontaminate": true,
+      "doc_to_decontamination_query": "{{text}}",
+      "metadata": {
+        "version": 1.0
+      }
+    }
+  },
+  "versions": {
+    "lambada_openai": 1.0
+  },
+  "n-shot": {
+    "lambada_openai": 0
+  },
+  "n-samples": {
+    "lambada_openai": {
+      "original": 5153,
+      "effective": 5153
+    }
+  },
+  "config": {
+    "model": "hf",
+    "model_args": "pretrained=EleutherAI/pythia-14m,revision=step140000,",
+    "model_num_parameters": 14067712,
+    "model_dtype": "torch.float16",
+    "model_revision": "step140000",
+    "model_sha": "5bb0fea58b8b2712c579d791ff1c8bf2790cca10",
+    "batch_size": "128",
+    "batch_sizes": [],
+    "device": "cuda",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": null,
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "51a7ca9",
+  "date": 1723764270.5154667,
+  "pretty_env_info": "PyTorch version: 2.3.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: CentOS Linux release 7.9.2009 (Core) (x86_64)\nGCC version: (GCC) 12.1.0\nClang version: Could not collect\nCMake version: Could not collect\nLibc version: glibc-2.17\n\nPython version: 3.12.1 (main, Jan 12 2024, 16:49:08) [GCC 12.1.0] (64-bit runtime)\nPython platform: Linux-3.10.0-1160.119.1.el7.tuxcare.els2.x86_64-x86_64-with-glibc2.17\nIs CUDA available: True\nCUDA runtime version: 12.4.99\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: GPU 0: Tesla V100S-PCIE-32GB\nNvidia driver version: 550.90.07\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture:          x86_64\nCPU op-mode(s):        32-bit, 64-bit\nByte Order:            Little Endian\nCPU(s):                32\nOn-line CPU(s) list:   0-31\nThread(s) per core:    1\nCore(s) per socket:    32\nSocket(s):             1\nNUMA node(s):          2\nVendor ID:             AuthenticAMD\nCPU family:            23\nModel:                 49\nModel name:            AMD EPYC 7502P 32-Core Processor\nStepping:              0\nCPU MHz:               1500.000\nCPU max MHz:           2500.0000\nCPU min MHz:           1500.0000\nBogoMIPS:              5000.08\nVirtualization:        AMD-V\nL1d cache:             32K\nL1i cache:             32K\nL2 cache:              512K\nL3 cache:              16384K\nNUMA node0 CPU(s):     0-15\nNUMA node1 CPU(s):     16-31\nFlags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc art rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_l2 cpb cat_l3 cdp_l3 hw_pstate sme ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip overflow_recov succor smca\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] torch==2.3.0\n[conda] Could not collect",
+  "transformers_version": "4.40.2",
+  "upper_git_hash": null,
+  "task_hashes": {},
+  "model_source": "hf",
+  "model_name": "EleutherAI/pythia-14m",
+  "model_name_sanitized": "EleutherAI__pythia-14m",
+  "start_time": 1146839.438632164,
+  "end_time": 1146918.215870824,
+  "total_evaluation_time_seconds": "78.77723866002634"
+}