End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25a029f5fdd2d54afbe54a04b8fa325002c96bb31ff5fe5693e87ba5496a484c
+oid sha256:d3cd7783a7cef04e08ab989781baac624a581ca2238ef68b92be07d6c5dfaa83
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.475056295585091e-07,
-    "train_runtime": 560.6623,
+    "train_loss": 6.273632436659682e-07,
+    "train_runtime": 590.7803,
     "train_samples": 11,
-    "train_samples_per_second": 0.571,
-    "train_steps_per_second": 0.036
+    "train_samples_per_second": 0.542,
+    "train_steps_per_second": 0.034
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.475056295585091e-07,
-    "train_runtime": 560.6623,
+    "train_loss": 6.273632436659682e-07,
+    "train_runtime": 590.7803,
     "train_samples": 11,
-    "train_samples_per_second": 0.571,
-    "train_steps_per_second": 0.036
+    "train_samples_per_second": 0.542,
+    "train_steps_per_second": 0.034
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 445.0625,
+      "completion_length": 485.375,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.5924282073974609,
-      "kl": 0.0007214371180452872,
+      "grad_norm": 0.46356967091560364,
+      "kl": 0.0006885775619593915,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 2.150634866207838,
-      "reward_std": 2.6777368709445,
-      "rewards/concensus_correctness_reward_func": 1.252875000005588,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 1.2886099740862846,
+      "reward_std": 1.0936285629868507,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5993848666548729,
-      "rewards/soft_format_reward_func": 0.03125,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": -0.18599999696016312,
+      "rewards/final_correctness_reward_func": 0.125,
+      "rewards/question_recreation_reward_func": 0.6098599601536989,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.3037500036880374,
       "step": 2
     },
     {
-      "completion_length": 413.5,
+      "completion_length": 412.67857142857144,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.3700309097766876,
-      "kl": 0.0007156268402468413,
+      "grad_norm": 0.4725421071052551,
+      "kl": 0.0007597233883903495,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.6294338777661324,
-      "reward_std": 1.0002400534493583,
-      "rewards/concensus_correctness_reward_func": 0.026857143800173486,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 0.05788119988782065,
+      "reward_std": 0.9135433307715825,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5103981580053057,
+      "rewards/question_recreation_reward_func": 0.36470261961221695,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.12210715562105179,
+      "rewards/xmlcount_reward_func": -0.37825000498976025,
       "step": 4
     },
     {
-      "completion_length": 437.75,
+      "completion_length": 445.67857142857144,
       "epoch": 2.0,
-      "grad_norm": 0.4982166886329651,
-      "kl": 0.0007187698169478349,
+      "grad_norm": 0.3603663146495819,
+      "kl": 0.0006844874088918525,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 0.9528258102280753,
-      "reward_std": 0.9797802226884025,
-      "rewards/concensus_correctness_reward_func": 0.008928571428571428,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "reward": 0.8103399872779846,
+      "reward_std": 1.038728028535843,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.508897245994636,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5441256782838276,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.14928572092737472,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": -0.037357138735907416,
       "step": 6
     },
     {
-      "completion_length": 419.125,
+      "completion_length": 454.59375,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.48654037714004517,
-      "kl": 0.000645398860797286,
+      "grad_norm": 0.28744569420814514,
+      "kl": 0.0006660163635388017,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.0086698783561587,
-      "reward_std": 0.7541907243430614,
-      "rewards/concensus_correctness_reward_func": 0.005750000011175871,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.7709572687745094,
+      "reward_std": 1.0267691239714622,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4138261363841593,
+      "rewards/question_recreation_reward_func": 0.43351975828409195,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.27659375593066216,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.07181250862777233,
       "step": 8
     },
     {
-      "completion_length": 378.67857142857144,
+      "completion_length": 441.07142857142856,
       "epoch": 3.3636363636363638,
-      "grad_norm": 4.351057529449463,
-      "kl": 0.0008367412457508701,
+      "grad_norm": 0.5901888608932495,
+      "kl": 0.0007456234972258764,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 5.878844491073063,
-      "reward_std": 2.495412162372044,
-      "rewards/concensus_correctness_reward_func": 4.309285715222359,
-      "rewards/consensus_reward_func": 0.5714285714285714,
+      "reward": 0.8694685186658587,
+      "reward_std": 1.0195827228682381,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.21428571428571427,
-      "rewards/question_recreation_reward_func": 0.5694876781531742,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6238256778035846,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.21435714619500296,
+      "rewards/xmlcount_reward_func": -0.0400714459163802,
       "step": 10
     },
     {
-      "completion_length": 462.4642857142857,
+      "completion_length": 447.64285714285717,
       "epoch": 4.0,
-      "grad_norm": 0.4533748924732208,
-      "kl": 0.0005763142731406593,
+      "grad_norm": 0.2877914607524872,
+      "kl": 0.0006377353607344308,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 0.9111057264464242,
-      "reward_std": 0.8011125581605094,
-      "rewards/concensus_correctness_reward_func": 0.015500000012772424,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 0.9620583398001534,
+      "reward_std": 1.000628330877849,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5330343118735722,
+      "rewards/question_recreation_reward_func": 0.627058357000351,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.07685714960098267,
+      "rewards/xmlcount_reward_func": -0.022142855184418813,
       "step": 12
     },
     {
-      "completion_length": 452.5,
+      "completion_length": 460.875,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.4300115704536438,
-      "kl": 0.0006732182373525575,
+      "grad_norm": 0.3677603006362915,
+      "kl": 0.0006959485108382069,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 3.085867512971163,
-      "reward_std": 2.1538953632116318,
-      "rewards/concensus_correctness_reward_func": 1.8807500000111759,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 0.7318503968417645,
+      "reward_std": 0.8866724520921707,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.5258361911401153,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.48185040429234505,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.05428127199411392,
+      "rewards/xmlcount_reward_func": 0.0625000037252903,
       "step": 14
     },
     {
-      "completion_length": 455.82142857142856,
+      "completion_length": 508.2142857142857,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.6953038573265076,
-      "kl": 0.0006783095803777022,
+      "grad_norm": 0.4596138000488281,
+      "kl": 0.0007280998480772334,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.9437104037829808,
-      "reward_std": 1.02949720621109,
-      "rewards/concensus_correctness_reward_func": 0.023714284811701094,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "reward": 0.8560242014271873,
+      "reward_std": 0.8382286131381989,
+      "rewards/concensus_correctness_reward_func": 0.033285713621548245,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.435281831238951,
-      "rewards/soft_format_reward_func": 0.017857142857142856,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5611670400415149,
+      "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.25257141462394167,
+      "rewards/xmlcount_reward_func": 0.04728572656001363,
       "step": 16
     },
     {
-      "completion_length": 394.60714285714283,
+      "completion_length": 382.7857142857143,
       "epoch": 6.0,
-      "grad_norm": 0.44728559255599976,
-      "kl": 0.0006316021691808211,
+      "grad_norm": 0.24859537184238434,
+      "kl": 0.0007194839459511318,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 3.7249111235141754,
-      "reward_std": 2.481673768588475,
-      "rewards/concensus_correctness_reward_func": 2.173000000949417,
-      "rewards/consensus_reward_func": 0.6428571428571429,
+      "reward": 0.8212329830442157,
+      "reward_std": 0.8151673674583435,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5788753777742386,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5104115701147488,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2587499873978751,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.2036785834601947,
       "step": 18
     },
     {
-      "completion_length": 373.5625,
+      "completion_length": 432.375,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.5277024507522583,
-      "kl": 0.0007235171324282419,
+      "grad_norm": 0.4400365352630615,
+      "kl": 0.0006472563909483142,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 3.255812767893076,
-      "reward_std": 2.323829635977745,
-      "rewards/concensus_correctness_reward_func": 1.9013750008307397,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 0.8123836005106568,
+      "reward_std": 0.9522764729335904,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.5847501866519451,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5303210988640785,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.1446874844841659,
+      "rewards/xmlcount_reward_func": 0.09456250071525574,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 6.475056295585091e-07,
-      "train_runtime": 560.6623,
-      "train_samples_per_second": 0.571,
-      "train_steps_per_second": 0.036
+      "train_loss": 6.273632436659682e-07,
+      "train_runtime": 590.7803,
+      "train_samples_per_second": 0.542,
+      "train_steps_per_second": 0.034
     }
   ],
   "logging_steps": 2,