End of training

adapter_config.json

@@ -25,12 +25,12 @@
   "revision": null,
   "target_modules": [
     "q_proj",
-    "k_proj",
-    "up_proj",
-    "o_proj",
-    "gate_proj",
     "down_proj",
-    "v_proj"
+    "o_proj",
+    "up_proj",
+    "k_proj",
+    "v_proj",
+    "gate_proj"
   ],
   "task_type": "CAUSAL_LM",
   "trainable_token_indices": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:254f6a79f845afb855ea6c1c2f0e95bdf7286d2bbb4f66d8add71faa72bcfcfa
+oid sha256:e5c55c8f367198bb94caf3958dc370442d6c19c6d1a2028cda0355678e80ae83
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.183008335687191e-07,
-    "train_runtime": 1688.5833,
-    "train_samples": 11,
-    "train_samples_per_second": 0.19,
-    "train_steps_per_second": 0.012
+    "train_loss": 1.1366962233694267e-07,
+    "train_runtime": 500.2968,
+    "train_samples": 1,
+    "train_samples_per_second": 0.64,
+    "train_steps_per_second": 0.04
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.183008335687191e-07,
-    "train_runtime": 1688.5833,
-    "train_samples": 11,
-    "train_samples_per_second": 0.19,
-    "train_steps_per_second": 0.012
+    "train_loss": 1.1366962233694267e-07,
+    "train_runtime": 500.2968,
+    "train_samples": 1,
+    "train_samples_per_second": 0.64,
+    "train_steps_per_second": 0.04
 }

trainer_state.json

@@ -2,7 +2,7 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 6.7272727272727275,
+  "epoch": 20.0,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -10,209 +10,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 559.25,
-      "epoch": 0.7272727272727273,
-      "grad_norm": 0.3945791721343994,
-      "kl": 0.0006233986823644955,
+      "completion_length": 745.0,
+      "epoch": 2.0,
+      "grad_norm": 0.18274663388729095,
+      "kl": 0.0,
       "learning_rate": 5e-07,
-      "loss": 0.0,
-      "reward": 0.8096773698925972,
-      "reward_std": 1.1500645354390144,
-      "rewards/concensus_correctness_reward_func": 0.054249998182058334,
-      "rewards/consensus_reward_func": 0.1875,
+      "loss": -0.0,
+      "reward": 1.084837019443512,
+      "reward_std": 1.1805158257484436,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.424021128565073,
+      "rewards/question_recreation_reward_func": 0.3814619928598404,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.09703124687075615,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": 0.6408749967813492,
       "step": 2
     },
     {
-      "completion_length": 536.5,
-      "epoch": 1.3636363636363638,
-      "grad_norm": 0.3390505909919739,
-      "kl": 0.0006377625561851476,
+      "completion_length": 778.75,
+      "epoch": 4.0,
+      "grad_norm": 0.12910106778144836,
+      "kl": 0.0005327967810444534,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.7590592705777713,
-      "reward_std": 0.9637055354458945,
-      "rewards/concensus_correctness_reward_func": 0.03057142879281725,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "reward": 0.07697300612926483,
+      "reward_std": 1.4115747213363647,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6418450006416866,
+      "rewards/question_recreation_reward_func": 0.4600980132818222,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.05621427749948842,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": -0.4456250071525574,
       "step": 4
     },
     {
-      "completion_length": 421.14285714285717,
-      "epoch": 2.0,
-      "grad_norm": 5.132058143615723,
-      "kl": 0.0009087596845347434,
+      "completion_length": 703.125,
+      "epoch": 6.0,
+      "grad_norm": 0.1341792196035385,
+      "kl": 0.0004545120464172214,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.349709503352642,
-      "reward_std": 1.2112282855170113,
-      "rewards/concensus_correctness_reward_func": 0.1360714307853154,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 0.26113903522491455,
+      "reward_std": 0.8645140528678894,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.42778094325746807,
+      "rewards/question_recreation_reward_func": 0.45626404881477356,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.48228571244648527,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": -0.19512498378753662,
       "step": 6
     },
     {
-      "completion_length": 451.59375,
-      "epoch": 2.7272727272727275,
-      "grad_norm": 0.37581542134284973,
-      "kl": 0.0007176933504524641,
+      "completion_length": 705.75,
+      "epoch": 8.0,
+      "grad_norm": 0.13170865178108215,
+      "kl": 0.0005088421748951077,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 0.9974680133163929,
-      "reward_std": 1.0404061805456877,
-      "rewards/concensus_correctness_reward_func": 0.05350000038743019,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 0.266592875123024,
+      "reward_std": 1.457129716873169,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4542492823675275,
+      "rewards/question_recreation_reward_func": 0.30996787548065186,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2397187538444996,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": -0.10587500035762787,
       "step": 8
     },
     {
-      "completion_length": 449.32142857142856,
-      "epoch": 3.3636363636363638,
-      "grad_norm": 0.31820976734161377,
-      "kl": 0.0007214340980031661,
+      "completion_length": 679.375,
+      "epoch": 10.0,
+      "grad_norm": 0.1503751128911972,
+      "kl": 0.0005099998670630157,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 0.8232330467019763,
-      "reward_std": 0.8050129456179482,
+      "reward": 0.0006678104400634766,
+      "reward_std": 0.6963486075401306,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5601258831364768,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.22504279017448425,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.04882141734872546,
+      "rewards/xmlcount_reward_func": -0.22437500953674316,
       "step": 10
     },
     {
-      "completion_length": 586.8928571428571,
-      "epoch": 4.0,
-      "grad_norm": 0.33116215467453003,
-      "kl": 0.0006629877191569124,
+      "completion_length": 757.5,
+      "epoch": 12.0,
+      "grad_norm": 0.1588054895401001,
+      "kl": 0.00046218409261200577,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.0386820648397719,
-      "reward_std": 1.0133664863450187,
-      "rewards/concensus_correctness_reward_func": 0.03128571382590702,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": -0.19009771943092346,
+      "reward_std": 1.6218498945236206,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4966463308249201,
+      "rewards/question_recreation_reward_func": 0.17290228977799416,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.22503571318728582,
+      "rewards/xmlcount_reward_func": -0.3630000352859497,
       "step": 12
     },
     {
-      "completion_length": 552.9375,
-      "epoch": 4.7272727272727275,
-      "grad_norm": 0.5910645723342896,
-      "kl": 0.0006863362868898548,
+      "completion_length": 654.125,
+      "epoch": 14.0,
+      "grad_norm": 0.1310453712940216,
+      "kl": 0.00050046292017214,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.0580328777432442,
-      "reward_std": 1.251479398459196,
-      "rewards/concensus_correctness_reward_func": 0.0702500008046627,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 0.7278059124946594,
+      "reward_std": 1.1783943176269531,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.572251632809639,
+      "rewards/question_recreation_reward_func": 0.2610559053719044,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.04053124599158764,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": 0.40424999594688416,
       "step": 14
     },
     {
-      "completion_length": 593.1428571428571,
-      "epoch": 5.363636363636363,
-      "grad_norm": 9.025611877441406,
-      "kl": 0.001219750677202163,
+      "completion_length": 703.5,
+      "epoch": 16.0,
+      "grad_norm": 0.1375514566898346,
+      "kl": 0.00047489034477621317,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 1.739785841533116,
-      "reward_std": 1.0538738071918488,
-      "rewards/concensus_correctness_reward_func": 0.09171428850718907,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": -0.11608892679214478,
+      "reward_std": 0.8753925263881683,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5370357962591308,
+      "rewards/question_recreation_reward_func": 0.5065360516309738,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.6110357151233724,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": -0.6851249933242798,
       "step": 16
     },
     {
-      "completion_length": 436.9642857142857,
-      "epoch": 6.0,
-      "grad_norm": 0.33825138211250305,
-      "kl": 0.0007806372845412365,
+      "completion_length": 646.875,
+      "epoch": 18.0,
+      "grad_norm": 0.11112207174301147,
+      "kl": 0.0005517661338672042,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 0.5506234424454826,
-      "reward_std": 0.7747731081077031,
-      "rewards/concensus_correctness_reward_func": 0.002999999959553991,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "reward": 0.1678205281496048,
+      "reward_std": 1.0244586765766144,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4423377215862274,
+      "rewards/question_recreation_reward_func": 0.29194553941488266,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.03757143632641861,
+      "rewards/xmlcount_reward_func": -0.12412500381469727,
       "step": 18
     },
     {
-      "completion_length": 497.59375,
-      "epoch": 6.7272727272727275,
-      "grad_norm": 0.5470736026763916,
-      "kl": 0.0007561461679870263,
+      "completion_length": 705.0,
+      "epoch": 20.0,
+      "grad_norm": 0.06440959125757217,
+      "kl": 0.0005306277598720044,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.2641245424747467,
-      "reward_std": 1.0695165917277336,
-      "rewards/concensus_correctness_reward_func": 0.02974999905563891,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 0.6555706560611725,
+      "reward_std": 0.8972451090812683,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.4777182573452592,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.47207067906856537,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.2410312513820827,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.1834999993443489,
       "step": 20
     },
     {
-      "epoch": 6.7272727272727275,
+      "epoch": 20.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 7.183008335687191e-07,
-      "train_runtime": 1688.5833,
-      "train_samples_per_second": 0.19,
-      "train_steps_per_second": 0.012
+      "train_loss": 1.1366962233694267e-07,
+      "train_runtime": 500.2968,
+      "train_samples_per_second": 0.64,
+      "train_steps_per_second": 0.04
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 20,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8550070dcbf31eb2b1a5f94205e0051e1f7447528d586fac90e792ff1b94288
+oid sha256:12d2a431517f57b1c782f600008325e94fca553837147faaf75b11b806c7cbdb
 size 5944