End of training

adapter_config.json

@@ -24,13 +24,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "v_proj",
     "k_proj",
+    "q_proj",
     "o_proj",
     "gate_proj",
-    "up_proj",
-    "down_proj"
+    "v_proj",
+    "down_proj",
+    "up_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 sha256:076bad1d1ad90a3408af3038dcb53d8424782f8dac2f21420050f842fcb4f033
+oid sha256:98789e1c319ee8007fe1e99225a88ffb524684a3240a3bf8bf3e2830d6ae50a6
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.104203099572714e-07,
-    "train_runtime": 1721.2094,
-    "train_samples": 100,
-    "train_samples_per_second": 0.186,
+    "train_loss": 5.854563340790264e-07,
+    "train_runtime": 1708.0615,
+    "train_samples": 11,
+    "train_samples_per_second": 0.187,
     "train_steps_per_second": 0.012
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.104203099572714e-07,
-    "train_runtime": 1721.2094,
-    "train_samples": 100,
-    "train_samples_per_second": 0.186,
+    "train_loss": 5.854563340790264e-07,
+    "train_runtime": 1708.0615,
+    "train_samples": 11,
+    "train_samples_per_second": 0.187,
     "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": 0.8,
+  "epoch": 6.7272727272727275,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -10,209 +10,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 473.9375,
-      "epoch": 0.08,
-      "grad_norm": 0.34969469904899597,
-      "kl": 0.0,
+      "completion_length": 542.625,
+      "epoch": 0.7272727272727273,
+      "grad_norm": 0.629473865032196,
+      "kl": 0.0006807459431001917,
       "learning_rate": 5e-07,
-      "loss": -0.0,
-      "reward": 1.483535647392273,
-      "reward_std": 1.2373702749609947,
-      "rewards/concensus_correctness_reward_func": 0.006062500178813934,
-      "rewards/consensus_reward_func": 0.5,
+      "loss": 0.0,
+      "reward": 1.040719661861658,
+      "reward_std": 1.0590924993157387,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5246606338769197,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.4371874928474426,
+      "rewards/question_recreation_reward_func": 0.47781341057270765,
+      "rewards/soft_format_reward_func": 0.015625,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.37540625780820847,
       "step": 2
     },
     {
-      "completion_length": 466.75,
-      "epoch": 0.16,
-      "grad_norm": 0.31218668818473816,
-      "kl": 0.0007654730725334957,
+      "completion_length": 505.75,
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.30012160539627075,
+      "kl": 0.0005313220857975207,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.7086614593863487,
-      "reward_std": 0.8326445110142231,
-      "rewards/concensus_correctness_reward_func": 0.06012500077486038,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 0.8662035976137433,
+      "reward_std": 1.3470306141035897,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5004427116364241,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.6393464761120933,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.10190625116229057,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": -0.005285704774515969,
       "step": 4
     },
     {
-      "completion_length": 488.75,
-      "epoch": 0.24,
-      "grad_norm": 0.3539212942123413,
-      "kl": 0.0007989688565430697,
+      "completion_length": 538.8571428571429,
+      "epoch": 2.0,
+      "grad_norm": 0.24394696950912476,
+      "kl": 0.0006991969047313822,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 0.9597348868846893,
-      "reward_std": 0.8448215201497078,
-      "rewards/concensus_correctness_reward_func": 0.054749999195337296,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 0.7815562805959156,
+      "reward_std": 1.1906610173838479,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5574536551721394,
+      "rewards/question_recreation_reward_func": 0.4404134175607136,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2850312504451722,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.10899999844176429,
       "step": 6
     },
     {
-      "completion_length": 429.4375,
-      "epoch": 0.32,
-      "grad_norm": 0.5844979286193848,
-      "kl": 0.000885855064552743,
+      "completion_length": 493.375,
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.43420788645744324,
+      "kl": 0.0006392400246113539,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 0.6777178943157196,
-      "reward_std": 0.7672723643481731,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 1.87135736271739,
+      "reward_std": 2.636794038116932,
+      "rewards/concensus_correctness_reward_func": 0.625,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.3103741565719247,
+      "rewards/question_recreation_reward_func": 0.5545448195189238,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.28921874798834324,
+      "rewards/xmlcount_reward_func": 0.17618749476969242,
       "step": 8
     },
     {
-      "completion_length": 436.5625,
-      "epoch": 0.4,
-      "grad_norm": 0.5599504113197327,
-      "kl": 0.0008282829003292136,
+      "completion_length": 528.4285714285714,
+      "epoch": 3.3636363636363638,
+      "grad_norm": 0.3044869899749756,
+      "kl": 0.0006325372461495655,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.1772124115377665,
-      "reward_std": 1.0518584689125419,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 1.6070085785218648,
+      "reward_std": 2.63333421094077,
+      "rewards/concensus_correctness_reward_func": 0.7142857142857143,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.5402123816311359,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.45354423991271425,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.3088750019669533,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.1534642885838236,
       "step": 10
     },
     {
-      "completion_length": 476.3125,
-      "epoch": 0.48,
-      "grad_norm": 1.196959376335144,
-      "kl": 0.0007865805309847929,
+      "completion_length": 571.8214285714286,
+      "epoch": 4.0,
+      "grad_norm": 0.27006351947784424,
+      "kl": 0.0005691683950967022,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 0.3430284606292844,
-      "reward_std": 1.131084494292736,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 1.1789613664150238,
+      "reward_std": 1.1718342304229736,
+      "rewards/concensus_correctness_reward_func": 0.04235714248248509,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.3210597266443074,
+      "rewards/question_recreation_reward_func": 0.4902113207748958,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": -0.05615623015910387,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.2892500066331455,
       "step": 12
     },
     {
-      "completion_length": 491.0625,
-      "epoch": 0.56,
-      "grad_norm": 0.3591214418411255,
-      "kl": 0.0007238566759042442,
+      "completion_length": 504.625,
+      "epoch": 4.7272727272727275,
+      "grad_norm": 0.3567564785480499,
+      "kl": 0.0006693297946185339,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 0.9437421336770058,
-      "reward_std": 1.0063903480768204,
-      "rewards/concensus_correctness_reward_func": 0.0,
+      "reward": 1.6142442487180233,
+      "reward_std": 2.231844250112772,
+      "rewards/concensus_correctness_reward_func": 0.6607500314712524,
       "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6040233373641968,
+      "rewards/question_recreation_reward_func": 0.6251505129039288,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.011593742296099663,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.015843747183680534,
       "step": 14
     },
     {
-      "completion_length": 481.4375,
-      "epoch": 0.64,
-      "grad_norm": 0.5718751549720764,
-      "kl": 0.0007593812551931478,
+      "completion_length": 531.7142857142857,
+      "epoch": 5.363636363636363,
+      "grad_norm": 1.1097288131713867,
+      "kl": 0.0006887656387074717,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 1.145691767334938,
-      "reward_std": 0.8273868970572948,
-      "rewards/concensus_correctness_reward_func": 0.012000000104308128,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 1.1217594317027502,
+      "reward_std": 1.284791009766715,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5370667604729533,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.4043308688061578,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.471624992787838,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.27099999359675814,
       "step": 16
     },
     {
-      "completion_length": 464.34375,
-      "epoch": 0.72,
-      "grad_norm": 0.41601094603538513,
-      "kl": 0.0008069156974670477,
+      "completion_length": 540.3571428571429,
+      "epoch": 6.0,
+      "grad_norm": 0.23262006044387817,
+      "kl": 0.0005829014075321279,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.989693470299244,
-      "reward_std": 1.9797653332352638,
-      "rewards/concensus_correctness_reward_func": 0.625,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.8904932481901986,
+      "reward_std": 0.9287492207118443,
+      "rewards/concensus_correctness_reward_func": 0.03550000063010624,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.25,
-      "rewards/question_recreation_reward_func": 0.5563809499144554,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.23018749989569187,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.5261361173221043,
+      "rewards/soft_format_reward_func": 0.017857142857142856,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.07885714939662389,
       "step": 18
     },
     {
-      "completion_length": 481.75,
-      "epoch": 0.8,
-      "grad_norm": 0.4525759518146515,
-      "kl": 0.0008362083317479119,
+      "completion_length": 575.96875,
+      "epoch": 6.7272727272727275,
+      "grad_norm": 0.3548136055469513,
+      "kl": 0.0005793773343611974,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.3046186864376068,
-      "reward_std": 1.2233120650053024,
-      "rewards/concensus_correctness_reward_func": 0.013875000178813934,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 1.8278657123446465,
+      "reward_std": 2.4451907724142075,
+      "rewards/concensus_correctness_reward_func": 0.6607500314712524,
+      "rewards/consensus_reward_func": 0.4375,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.6678999401628971,
+      "rewards/question_recreation_reward_func": 0.5390531914308667,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.23221874982118607,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.09681251272559166,
       "step": 20
     },
     {
-      "epoch": 0.8,
+      "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 7.104203099572714e-07,
-      "train_runtime": 1721.2094,
-      "train_samples_per_second": 0.186,
+      "train_loss": 5.854563340790264e-07,
+      "train_runtime": 1708.0615,
+      "train_samples_per_second": 0.187,
       "train_steps_per_second": 0.012
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
+  "num_train_epochs": 10,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6eddf13391de541790c176fc7331830e211a0c4f6babb2091a46b97c829559a8
+oid sha256:054f539369d59b030ad14234866b4883dde23295313aeff1f42632bd5c104251
 size 5944