{ "run_info": { "created_at": "2025-06-20T08:01:25+00:00", "total_time": 2714.5956150429993, "experiment_name": "prompt_tuning/llama-3.2-3B-lr_0.001", "peft_branch": "main", "train_config": { "model_id": "meta-llama/Llama-3.2-3B", "dtype": "bfloat16", "max_seq_length": 768, "batch_size": 4, "batch_size_eval": 50, "max_steps": 5000, "eval_steps": 250, "compile": false, "query_template": "Question: {query} Think step by step.\nAnswer:", "seed": 0, "grad_norm_clip": 1.0, "optimizer_type": "AdamW", "optimizer_kwargs": { "lr": 0.001 }, "lr_scheduler": "cosine", "use_amp": false, "autocast_adapter_dtype": true, "generation_kwargs": { "max_length": 800, "max_new_tokens": 300 }, "attn_implementation": null }, "peft_config": { "task_type": "CAUSAL_LM", "peft_type": "PROMPT_TUNING", "auto_mapping": null, "base_model_name_or_path": "meta-llama/Llama-3.2-3B", "revision": null, "inference_mode": false, "num_virtual_tokens": 200, "token_dim": 3072, "num_transformer_submodules": 1, "num_attention_heads": 24, "num_layers": 28, "prompt_tuning_init": "RANDOM", "prompt_tuning_init_text": null, "tokenizer_name_or_path": null, "tokenizer_kwargs": null }, "error_msg": "" }, "train_info": { "accelerator_memory_reserved_avg": 15297364466, "accelerator_memory_max": 24408752128, "accelerator_memory_reserved_99th": 20650676715, "train_time": 2394.4007484640024, "file_size": 2457728, "num_trainable_params": 614400, "num_total_params": 3213364224, "status": "success", "metrics": [ { "step": 250, "valid accuracy": 0.0, "train loss": 2.454602773666382, "train samples": 1000, "train time": 46.58359175696387, "eval time": 15.906975480989786, "tokens / sec": 4544.926486231061, "mem allocated avg": 7082736850.944, "mem reserved avg": 15330147565.568, "elapsed time": 120.51601758999459 }, { "step": 500, "valid accuracy": 0.02, "train loss": 1.4034885478019714, "train samples": 2000, "train time": 45.99672341402038, "eval time": 15.859127072995761, "tokens / sec": 4521.952534049426, "mem allocated avg": 7075398952.96, "mem reserved avg": 15237637996.544, "elapsed time": 234.56530582100095 }, { "step": 750, "valid accuracy": 0.1, "train loss": 1.051814435005188, "train samples": 3000, "train time": 45.34941398198134, "eval time": 15.839738530004979, "tokens / sec": 4727.756792738001, "mem allocated avg": 7085216630.784, "mem reserved avg": 15378130403.328, "elapsed time": 347.9996997119888 }, { "step": 1000, "valid accuracy": 0.2, "train loss": 0.9425526282787323, "train samples": 4000, "train time": 44.85872337181354, "eval time": 15.849193180998554, "tokens / sec": 4644.269482954245, "mem allocated avg": 7077280739.328, "mem reserved avg": 15280109518.848, "elapsed time": 460.8599872249906 }, { "step": 1250, "valid accuracy": 0.2, "train loss": 0.9085307500362396, "train samples": 5000, "train time": 45.535731699026655, "eval time": 15.864107311004773, "tokens / sec": 4579.656287909338, "mem allocated avg": 7076838449.152, "mem reserved avg": 15263508463.616, "elapsed time": 574.5614464429964 }, { "step": 1500, "valid accuracy": 0.18, "train loss": 0.8753413548469543, "train samples": 6000, "train time": 45.47140344994841, "eval time": 15.851111587006017, "tokens / sec": 4603.5746451155, "mem allocated avg": 7078501443.584, "mem reserved avg": 15280914825.216, "elapsed time": 688.3081236659928 }, { "step": 1750, "valid accuracy": 0.18, "train loss": 0.8501973593235016, "train samples": 7000, "train time": 45.876367467062664, "eval time": 15.86328411300201, "tokens / sec": 4563.460700115549, "mem allocated avg": 7079126001.664, "mem reserved avg": 15302154780.672, "elapsed time": 802.3839824919996 }, { "step": 2000, "valid accuracy": 0.3, "train loss": 0.8353641645908356, "train samples": 8000, "train time": 45.395122604924836, "eval time": 15.847279680005158, "tokens / sec": 4575.293293237354, "mem allocated avg": 7075813670.912, "mem reserved avg": 15257200230.4, "elapsed time": 915.8055839799927 }, { "step": 2250, "valid accuracy": 0.26, "train loss": 0.8205823216438294, "train samples": 9000, "train time": 46.531550297062495, "eval time": 15.857669960998464, "tokens / sec": 4619.403364550472, "mem allocated avg": 7087054014.464, "mem reserved avg": 15417707855.872, "elapsed time": 1030.8605109579948 }, { "step": 2500, "valid accuracy": 0.24, "train loss": 0.8074139108657837, "train samples": 10000, "train time": 45.232053409854416, "eval time": 15.864067172005889, "tokens / sec": 4553.562893413265, "mem allocated avg": 7073174814.72, "mem reserved avg": 15210467295.232, "elapsed time": 1144.3065934619954 }, { "step": 2750, "valid accuracy": 0.22, "train loss": 0.800323983669281, "train samples": 11000, "train time": 46.27672885800712, "eval time": 15.85089660200174, "tokens / sec": 4578.564760921707, "mem allocated avg": 7083499849.728, "mem reserved avg": 15345020567.552, "elapsed time": 1258.9190711479896 }, { "step": 3000, "valid accuracy": 0.28, "train loss": 0.7878623747825623, "train samples": 12000, "train time": 45.57083585388318, "eval time": 15.872650785997394, "tokens / sec": 4580.3636490071885, "mem allocated avg": 7078042595.328, "mem reserved avg": 15285402730.496, "elapsed time": 1372.7267461329902 }, { "step": 3250, "valid accuracy": 0.3, "train loss": 0.7943042907714843, "train samples": 13000, "train time": 45.666222987070796, "eval time": 15.852009978989372, "tokens / sec": 4618.314942746877, "mem allocated avg": 7079504875.52, "mem reserved avg": 15299428483.072, "elapsed time": 1486.5100108069892 }, { "step": 3500, "valid accuracy": 0.28, "train loss": 0.780832305431366, "train samples": 14000, "train time": 45.84015418085619, "eval time": 15.86955204399419, "tokens / sec": 4575.6826901685245, "mem allocated avg": 7078824071.168, "mem reserved avg": 15300871323.648, "elapsed time": 1600.7413567879994 }, { "step": 3750, "valid accuracy": 0.32, "train loss": 0.7758122501373291, "train samples": 15000, "train time": 46.99727132692351, "eval time": 15.8490629579901, "tokens / sec": 4610.969826153641, "mem allocated avg": 7089586788.352, "mem reserved avg": 15444173914.112, "elapsed time": 1716.2785189549904 }, { "step": 4000, "valid accuracy": 0.36, "train loss": 0.7912874612808227, "train samples": 16000, "train time": 45.15887627698248, "eval time": 15.855249352011015, "tokens / sec": 4525.644056031772, "mem allocated avg": 7071318118.4, "mem reserved avg": 15188732411.904, "elapsed time": 1829.5188424160006 }, { "step": 4250, "valid accuracy": 0.36, "train loss": 0.7664959132671356, "train samples": 17000, "train time": 46.26589757904003, "eval time": 15.853440922001028, "tokens / sec": 4569.002463182864, "mem allocated avg": 7081992153.088, "mem reserved avg": 15327354159.104, "elapsed time": 1944.2481972599926 }, { "step": 4500, "valid accuracy": 0.34, "train loss": 0.7785169410705567, "train samples": 18000, "train time": 45.61058669183694, "eval time": 15.866839458991308, "tokens / sec": 4556.354457882774, "mem allocated avg": 7075963725.824, "mem reserved avg": 15250623561.728, "elapsed time": 2058.0909812989994 }, { "step": 4750, "valid accuracy": 0.32, "train loss": 0.7709811532497406, "train samples": 19000, "train time": 45.832340708962874, "eval time": 15.847010081997723, "tokens / sec": 4580.586475674911, "mem allocated avg": 7079141249.024, "mem reserved avg": 15295871713.28, "elapsed time": 2172.3217773149954 }, { "step": 5000, "valid accuracy": 0.3, "train loss": 0.7790318930149078, "train samples": 20000, "train time": 44.844002045996604, "eval time": 15.846091532002902, "tokens / sec": 4644.545323728393, "mem allocated avg": 7075675734.016, "mem reserved avg": 15251831521.28, "elapsed time": 2285.3788618499966 }, { "step": 5000, "test accuracy": 0.25246398786959817, "train loss": 0.7790318930149078, "train samples": 20000, "train total tokens": 4198051 } ] }, "meta_info": { "model_info": { "sha": "13afe5124825b4f3751f836b40dafda64c1ed062", "created_at": "2024-09-18T15:23:48+00:00" }, "dataset_info": { "metamath": { "sha": "aa4f34d3d2d3231299b5b03d9b3e5a20da45aa18", "created_at": "2023-09-21T17:22:46+00:00" }, "gsm8k": { "sha": "e53f048856ff4f594e959d75785d2c2d37b678ee", "created_at": "2022-04-12T10:22:10+00:00" } }, "package_info": { "transformers-version": "4.52.4", "transformers-commit-hash": null, "peft-version": "0.15.2.dev0", "peft-commit-hash": "5fe7f8f8abe914d313fc3751f2ea92de7718fbaf", "datasets-version": "3.6.0", "datasets-commit-hash": null, "bitsandbytes-version": "0.46.0", "bitsandbytes-commit-hash": null, "torch-version": "2.7.1+cu126", "torch-commit-hash": null }, "system_info": { "system": "Linux", "release": "6.8.0-1029-aws", "version": "#31-Ubuntu SMP Wed Apr 23 18:42:41 UTC 2025", "machine": "x86_64", "processor": "x86_64", "accelerator": "NVIDIA L40S" }, "pytorch_info": "PyTorch built with:\n - GCC 11.2\n - C++ Version: 201703\n - Intel(R) oneAPI Math Kernel Library Version 2024.2-Product Build 20240605 for Intel(R) 64 architecture applications\n - Intel(R) MKL-DNN v3.7.1 (Git Hash 8d263e693366ef8db40acc569cc7d8edf644556d)\n - OpenMP 201511 (a.k.a. OpenMP 4.5)\n - LAPACK is enabled (usually provided by MKL)\n - NNPACK is enabled\n - CPU capability usage: AVX2\n - CUDA Runtime 12.6\n - NVCC architecture flags: -gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_90,code=sm_90\n - CuDNN 90.7.1 (built against CUDA 12.8)\n - Built with CuDNN 90.5.1\n - Magma 2.6.1\n - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, COMMIT_SHA=e2d141dbde55c2a4370fac5165b0561b6af4798b, CUDA_VERSION=12.6, CUDNN_VERSION=9.5.1, CXX_COMPILER=/opt/rh/gcc-toolset-11/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=1 -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DLIBKINETO_NOXPUPTI=ON -DUSE_FBGEMM -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=range-loop-construct -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-unknown-pragmas -Wno-unused-parameter -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, TORCH_VERSION=2.7.1, USE_CUDA=ON, USE_CUDNN=ON, USE_CUSPARSELT=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_GLOO=ON, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF, \n" } }