End of training

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: Gensyn/Qwen2.5-0.5B-Instruct
+base_model: Gensyn/Qwen2.5-1.5B-Instruct
 library_name: transformers
 model_name: gensyn-checkpoints-beaked_tangled_caribou
 tags:
@@ -8,13 +8,14 @@ tags:
 - grpo
 - gensyn
 - I am beaked tangled caribou
+- unsloth
 - trl
 licence: license
 ---
 
 # Model Card for gensyn-checkpoints-beaked_tangled_caribou
 
-This model is a fine-tuned version of [Gensyn/Qwen2.5-0.5B-Instruct](https://huggingface.co/Gensyn/Qwen2.5-0.5B-Instruct).
+This model is a fine-tuned version of [Gensyn/Qwen2.5-1.5B-Instruct](https://huggingface.co/Gensyn/Qwen2.5-1.5B-Instruct).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -39,7 +40,7 @@ This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing
 
 - TRL: 0.15.2
 - Transformers: 4.51.3
-- Pytorch: 2.5.1
+- Pytorch: 2.6.0
 - Datasets: 3.5.0
 - Tokenizers: 0.21.1
 

adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Gensyn/Qwen2.5-1.5B-Instruct",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "gate_proj",
+    "k_proj",
+    "down_proj",
+    "v_proj",
+    "q_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29f7591262189b2502396df731d931c54aedda28f273757682035a8d5e6875cc
+size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.9904466477659298,
-    "train_runtime": 713.0205,
-    "train_samples": 28,
-    "train_samples_per_second": 0.449,
-    "train_steps_per_second": 0.028
+    "train_loss": 7.802889115282596e-07,
+    "train_runtime": 1687.4123,
+    "train_samples": 19,
+    "train_samples_per_second": 0.19,
+    "train_steps_per_second": 0.012
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.9904466477659298,
-    "train_runtime": 713.0205,
-    "train_samples": 28,
-    "train_samples_per_second": 0.449,
-    "train_steps_per_second": 0.028
+    "train_loss": 7.802889115282596e-07,
+    "train_runtime": 1687.4123,
+    "train_samples": 19,
+    "train_samples_per_second": 0.19,
+    "train_steps_per_second": 0.012
 }

trainer_state.json

@@ -2,7 +2,7 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 5.0,
+  "epoch": 4.0,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -10,209 +10,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 247.21875,
-      "epoch": 0.5714285714285714,
-      "grad_norm": 6.694652557373047,
-      "kl": 0.0,
+      "completion_length": 306.78125,
+      "epoch": 0.42105263157894735,
+      "grad_norm": 0.4840872585773468,
+      "kl": 0.0007103236821421888,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 3.429838713258505,
-      "reward_std": 0.7307246705458965,
-      "rewards/concensus_correctness_reward_func": 0.713249996304512,
-      "rewards/consensus_reward_func": 1.0,
+      "reward": 1.4722652584314346,
+      "reward_std": 1.0188412675634027,
+      "rewards/concensus_correctness_reward_func": 0.012500000186264515,
+      "rewards/consensus_reward_func": 0.4375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.6806199923157692,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6029215175658464,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.1875,
-      "rewards/xmlcount_reward_func": 0.7859687637537718,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.4193437611684203,
       "step": 2
     },
     {
-      "completion_length": 125.83333333333333,
-      "epoch": 1.0,
-      "grad_norm": 16.098657608032227,
-      "kl": 0.10447756666690111,
+      "completion_length": 354.25,
+      "epoch": 0.8421052631578947,
+      "grad_norm": 0.6972860097885132,
+      "kl": 0.0008639320731163025,
       "learning_rate": 4.864543104251586e-07,
-      "loss": 0.0001,
-      "reward": 5.87607725461324,
-      "reward_std": 0.4451955556869507,
-      "rewards/concensus_correctness_reward_func": 1.698999971151352,
-      "rewards/consensus_reward_func": 1.5833333333333333,
+      "loss": 0.0,
+      "reward": 0.7589982934296131,
+      "reward_std": 0.8806997202336788,
+      "rewards/concensus_correctness_reward_func": 0.010625000111758709,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.964202269911766,
+      "rewards/question_recreation_reward_func": 0.2897482868283987,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.3958333333333333,
-      "rewards/xmlcount_reward_func": 1.2337083319822948,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.271125003695488,
       "step": 4
     },
     {
-      "completion_length": 135.03125,
-      "epoch": 1.5714285714285714,
-      "grad_norm": 256478.375,
-      "kl": 8980.802092778962,
+      "completion_length": 429.85714285714283,
+      "epoch": 1.2105263157894737,
+      "grad_norm": 0.39067697525024414,
+      "kl": 0.0007936381651753825,
       "learning_rate": 4.472851273490984e-07,
-      "loss": 8.9808,
-      "reward": 5.556592166423798,
-      "reward_std": 0.5139178396202624,
-      "rewards/concensus_correctness_reward_func": 1.4853749834001064,
-      "rewards/consensus_reward_func": 1.5,
+      "loss": 0.0,
+      "reward": 1.0712779590061732,
+      "reward_std": 1.2960399900163924,
+      "rewards/concensus_correctness_reward_func": 0.0019285714785967553,
+      "rewards/consensus_reward_func": 0.42857142857142855,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.8682797141373158,
+      "rewards/question_recreation_reward_func": 0.4723136840122087,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.453125,
-      "rewards/xmlcount_reward_func": 1.2498124986886978,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.16846428332584246,
       "step": 6
     },
     {
-      "completion_length": 145.66666666666666,
-      "epoch": 2.0,
-      "grad_norm": 15.734909057617188,
-      "kl": 12.738230961724184,
+      "completion_length": 322.21875,
+      "epoch": 1.631578947368421,
+      "grad_norm": 9.73143482208252,
+      "kl": 0.001008568644465413,
       "learning_rate": 3.867370395306068e-07,
-      "loss": 0.0096,
-      "reward": 5.932813902695973,
-      "reward_std": 0.21585626168719804,
-      "rewards/concensus_correctness_reward_func": 1.4209999740123749,
-      "rewards/consensus_reward_func": 1.8333333333333333,
+      "loss": 0.0,
+      "reward": 1.2451195530593395,
+      "reward_std": 0.7887402325868607,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.08333333333333333,
-      "rewards/question_recreation_reward_func": 0.963147242863973,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.4551508268341422,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.3958333333333333,
-      "rewards/xmlcount_reward_func": 1.2361666758855183,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.41496875509619713,
       "step": 8
     },
     {
-      "completion_length": 151.0,
-      "epoch": 2.571428571428571,
-      "grad_norm": 27.214447021484375,
-      "kl": 4.368775271810591,
+      "completion_length": 394.0357142857143,
+      "epoch": 2.0,
+      "grad_norm": 0.4145382344722748,
+      "kl": 0.0007304587717434126,
       "learning_rate": 3.1137137178519977e-07,
-      "loss": 0.0044,
-      "reward": 5.704393148422241,
-      "reward_std": 0.6130428695760202,
-      "rewards/concensus_correctness_reward_func": 1.5178749822080135,
-      "rewards/consensus_reward_func": 1.625,
+      "loss": 0.0,
+      "reward": 1.0647335286651338,
+      "reward_std": 1.286596519606454,
+      "rewards/concensus_correctness_reward_func": 0.014285714498588018,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.8489869562909007,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.421875,
-      "rewards/xmlcount_reward_func": 1.228156253695488,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.4280192255973816,
+      "rewards/soft_format_reward_func": 0.017857142857142856,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": -0.0025714199457849774,
       "step": 10
     },
     {
-      "completion_length": 126.29166666666667,
-      "epoch": 3.0,
-      "grad_norm": 20.617341995239258,
-      "kl": 0.49174783130486804,
+      "completion_length": 315.5625,
+      "epoch": 2.4210526315789473,
+      "grad_norm": 0.48806658387184143,
+      "kl": 0.0008116730095935054,
       "learning_rate": 2.2935516363191693e-07,
-      "loss": 0.0004,
-      "reward": 5.857549011707306,
-      "reward_std": 0.021132514579221606,
-      "rewards/concensus_correctness_reward_func": 1.4578333099683125,
-      "rewards/consensus_reward_func": 1.6666666666666667,
+      "loss": 0.0,
+      "reward": 1.2011400386691093,
+      "reward_std": 0.9617396891117096,
+      "rewards/concensus_correctness_reward_func": 0.007125000120140612,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.9830490450064341,
+      "rewards/final_correctness_reward_func": 0.125,
+      "rewards/question_recreation_reward_func": 0.58567128283903,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.5,
-      "rewards/xmlcount_reward_func": 1.25,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.15521876234561205,
       "step": 12
     },
     {
-      "completion_length": 151.6875,
-      "epoch": 3.571428571428571,
-      "grad_norm": 42.12635040283203,
-      "kl": 2.118111965013668,
+      "completion_length": 347.15625,
+      "epoch": 2.8421052631578947,
+      "grad_norm": 0.5586909055709839,
+      "kl": 0.0008663672997499816,
       "learning_rate": 1.4957614383675767e-07,
-      "loss": 0.0021,
-      "reward": 5.8429388999938965,
-      "reward_std": 0.42098340444499627,
-      "rewards/concensus_correctness_reward_func": 1.5896249823272228,
-      "rewards/consensus_reward_func": 1.6875,
+      "loss": 0.0,
+      "reward": 0.9973086006939411,
+      "reward_std": 1.0326183773577213,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.9300950989127159,
+      "rewards/question_recreation_reward_func": 0.4399023484438658,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.40625,
-      "rewards/xmlcount_reward_func": 1.2294687554240227,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.22928125644102693,
       "step": 14
     },
     {
-      "completion_length": 147.45833333333334,
-      "epoch": 4.0,
-      "grad_norm": 17.08171844482422,
-      "kl": 16.20084721719225,
+      "completion_length": 377.39285714285717,
+      "epoch": 3.2105263157894735,
+      "grad_norm": 0.4769853353500366,
+      "kl": 0.0008522716650207128,
       "learning_rate": 8.067960709356478e-08,
-      "loss": 0.0122,
-      "reward": 5.538925488789876,
-      "reward_std": 0.7823425438255072,
-      "rewards/concensus_correctness_reward_func": 1.5709999750057857,
-      "rewards/consensus_reward_func": 1.5833333333333333,
+      "loss": 0.0,
+      "reward": 1.2226124150412423,
+      "reward_std": 1.2815379330090113,
+      "rewards/concensus_correctness_reward_func": 0.05085714374269758,
+      "rewards/consensus_reward_func": 0.5714285714285714,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.8435088787227869,
+      "rewards/question_recreation_reward_func": 0.39936240017414093,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.375,
-      "rewards/xmlcount_reward_func": 1.1660833358764648,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.18310714566281863,
       "step": 16
     },
     {
-      "completion_length": 126.90625,
-      "epoch": 4.571428571428571,
-      "grad_norm": 41542.9296875,
-      "kl": 894.772123008268,
+      "completion_length": 308.125,
+      "epoch": 3.6315789473684212,
+      "grad_norm": 6.139931678771973,
+      "kl": 0.0008203142569982447,
       "learning_rate": 3.013156219837776e-08,
-      "loss": 0.8948,
-      "reward": 5.760149195790291,
-      "reward_std": 0.5905774181082961,
-      "rewards/concensus_correctness_reward_func": 1.4777499809861183,
-      "rewards/consensus_reward_func": 1.6875,
+      "loss": 0.0,
+      "reward": 1.3216048330068588,
+      "reward_std": 1.2894400879740715,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.9594617262482643,
+      "rewards/final_correctness_reward_func": 0.125,
+      "rewards/question_recreation_reward_func": 0.5774798067286611,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.421875,
-      "rewards/xmlcount_reward_func": 1.2135625034570694,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.19725001230835915,
       "step": 18
     },
     {
-      "completion_length": 140.375,
-      "epoch": 5.0,
-      "grad_norm": 12.598287582397461,
-      "kl": 0.34355814941227436,
+      "completion_length": 361.89285714285717,
+      "epoch": 4.0,
+      "grad_norm": 0.6704761981964111,
+      "kl": 0.0007814069561261151,
       "learning_rate": 3.4096741493194193e-09,
-      "loss": 0.0003,
-      "reward": 5.7935870885849,
-      "reward_std": 0.2697240054840222,
-      "rewards/concensus_correctness_reward_func": 1.5494166413942974,
-      "rewards/consensus_reward_func": 1.5833333333333333,
+      "loss": 0.0,
+      "reward": 1.4310967241014754,
+      "reward_std": 0.652162033532347,
+      "rewards/concensus_correctness_reward_func": 0.006071428635290691,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.08333333333333333,
-      "rewards/question_recreation_reward_func": 0.9316704521576563,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.518061023737703,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.4166666666666667,
-      "rewards/xmlcount_reward_func": 1.2291666666666667,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.5319642871618271,
       "step": 20
     },
     {
-      "epoch": 5.0,
+      "epoch": 4.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 0.9904466477659298,
-      "train_runtime": 713.0205,
-      "train_samples_per_second": 0.449,
-      "train_steps_per_second": 0.028
+      "train_loss": 7.802889115282596e-07,
+      "train_runtime": 1687.4123,
+      "train_samples_per_second": 0.19,
+      "train_steps_per_second": 0.012
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 7,
+  "num_train_epochs": 5,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -227,7 +227,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 2,
+  "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e607b2b78e9388edc4729eca802c846773c3431bc6ae2d5141a01f3641540a39
-size 5880
+oid sha256:58ad81c61c18eda5e86440c8a7b0b33c341a7c6271456561ecb785686a740916
+size 5944