End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:efb82a21194f3090dbff069f121afe377a56fcf499039a7d46b74aa7c08ddca8
+oid sha256:1f402f194f59d4e1f5502e2249371400a0f9ff3a86aef6283c6f4f811c32d4ef
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.443942479312682e-07,
-    "train_runtime": 611.3719,
+    "train_loss": 7.208689851267991e-07,
+    "train_runtime": 624.2234,
     "train_samples": 11,
-    "train_samples_per_second": 0.523,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.513,
+    "train_steps_per_second": 0.032
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.443942479312682e-07,
-    "train_runtime": 611.3719,
+    "train_loss": 7.208689851267991e-07,
+    "train_runtime": 624.2234,
     "train_samples": 11,
-    "train_samples_per_second": 0.523,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.513,
+    "train_steps_per_second": 0.032
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 417.375,
+      "completion_length": 487.5,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.332813560962677,
-      "kl": 0.0007562078390037641,
+      "grad_norm": 0.3739618957042694,
+      "kl": 0.0006633192897425033,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 1.0843788953498006,
-      "reward_std": 0.9088983843103051,
-      "rewards/concensus_correctness_reward_func": 0.03125,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 1.2776643633842468,
+      "reward_std": 1.1137661971151829,
+      "rewards/concensus_correctness_reward_func": 0.0253749992698431,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.6132851466536522,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5647268481552601,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.17421874590218067,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.28131250478327274,
       "step": 2
     },
     {
-      "completion_length": 388.2857142857143,
+      "completion_length": 424.67857142857144,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.3792736530303955,
-      "kl": 0.0007836362804352705,
+      "grad_norm": 0.6920000910758972,
+      "kl": 0.0007849830851357963,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 3.0097297855785916,
-      "reward_std": 2.510497753109251,
-      "rewards/concensus_correctness_reward_func": 1.4622857230050224,
-      "rewards/consensus_reward_func": 0.42857142857142855,
+      "reward": 1.09174644947052,
+      "reward_std": 1.0946569357599532,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.2857142857142857,
-      "rewards/question_recreation_reward_func": 0.522158431155341,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.636567873614175,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.2931428523734212,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.24089284879820688,
       "step": 4
     },
     {
-      "completion_length": 467.92857142857144,
+      "completion_length": 424.9642857142857,
       "epoch": 2.0,
-      "grad_norm": 0.6137570738792419,
-      "kl": 0.0007571280335209199,
+      "grad_norm": 1.9624477624893188,
+      "kl": 0.000967897308458175,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.2021850666829519,
-      "reward_std": 1.1009544559887476,
-      "rewards/concensus_correctness_reward_func": 0.03571428571428571,
+      "reward": 0.9173746705055237,
+      "reward_std": 0.9211063001837049,
+      "rewards/concensus_correctness_reward_func": 0.028999999165534973,
       "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6813279262610844,
+      "rewards/question_recreation_reward_func": 0.4704460863556181,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.2529999954359872,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.1679285785981587,
       "step": 6
     },
     {
-      "completion_length": 460.34375,
+      "completion_length": 499.75,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.49027806520462036,
-      "kl": 0.0007044844314805232,
+      "grad_norm": 0.41693592071533203,
+      "kl": 0.0007434094113705214,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.1509894523769617,
-      "reward_std": 0.8515053279697895,
+      "reward": 1.1772202551364899,
+      "reward_std": 1.1608079262077808,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.5470832139253616,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6437514759600163,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.21328124403953552,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.2522187577560544,
       "step": 8
     },
     {
-      "completion_length": 434.2857142857143,
+      "completion_length": 459.7857142857143,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.5033062100410461,
-      "kl": 0.0007557787799409457,
+      "grad_norm": 0.4088006913661957,
+      "kl": 0.0006640334613621235,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.2265301176479884,
-      "reward_std": 0.7716207930019924,
+      "reward": 0.7122706749609539,
+      "reward_std": 0.8251376960958753,
       "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": 0.5335301075662885,
+      "rewards/question_recreation_reward_func": 0.5606278372662408,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.3179999887943268,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": -0.20549998033259595,
       "step": 10
     },
     {
-      "completion_length": 579.6785714285714,
+      "completion_length": 477.5357142857143,
       "epoch": 4.0,
-      "grad_norm": 0.33078819513320923,
-      "kl": 0.0015195429192057677,
+      "grad_norm": 0.4236905574798584,
+      "kl": 0.000740286594788943,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.4454642704554967,
-      "reward_std": 1.529354282787868,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 0.7668512889317104,
+      "reward_std": 1.0755319893360138,
+      "rewards/concensus_correctness_reward_func": 0.0012857142303671157,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.21428571428571427,
-      "rewards/question_recreation_reward_func": 0.5894285227571215,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.44456556013652254,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.28460715285369326,
+      "rewards/xmlcount_reward_func": 0.10671427845954895,
       "step": 12
     },
     {
-      "completion_length": 462.875,
+      "completion_length": 500.125,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.4308289587497711,
-      "kl": 0.0006815507949795574,
+      "grad_norm": 0.5966536998748779,
+      "kl": 0.0007182688859757036,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.7030750401318073,
-      "reward_std": 2.0401263535022736,
-      "rewards/concensus_correctness_reward_func": 0.65625,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 1.5318920835852623,
+      "reward_std": 1.2969727516174316,
+      "rewards/concensus_correctness_reward_func": 0.12087499862536788,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.5872938558459282,
-      "rewards/soft_format_reward_func": 0.015625,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6593920886516571,
+      "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.13140624482184649,
+      "rewards/xmlcount_reward_func": 0.5016250014305115,
       "step": 14
     },
     {
-      "completion_length": 378.25,
+      "completion_length": 517.6071428571429,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.43900179862976074,
-      "kl": 0.0007973652266498123,
+      "grad_norm": 0.6156628131866455,
+      "kl": 0.0008096090418153576,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 1.3655415688242232,
-      "reward_std": 1.3868229729788644,
-      "rewards/concensus_correctness_reward_func": 0.03571428571428571,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 1.1096807718276978,
+      "reward_std": 1.1326228763375963,
+      "rewards/concensus_correctness_reward_func": 0.054285713604518344,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6181844387735639,
+      "rewards/question_recreation_reward_func": 0.5367522026811328,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.05357142857142857,
-      "rewards/xmlcount_reward_func": 0.15807141150747026,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.30435714232070105,
       "step": 16
     },
     {
-      "completion_length": 573.0357142857143,
+      "completion_length": 463.7142857142857,
       "epoch": 6.0,
-      "grad_norm": 0.24938742816448212,
-      "kl": 0.0007200419869539994,
+      "grad_norm": 0.3831884562969208,
+      "kl": 0.0006995330830769879,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.0222688146999903,
-      "reward_std": 0.9925270250865391,
+      "reward": 1.0206448010035925,
+      "reward_std": 1.3927564195224218,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.4843402398484094,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.4904304998261588,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.23435714734452112,
+      "rewards/xmlcount_reward_func": 0.012357139161654882,
       "step": 18
     },
     {
-      "completion_length": 517.53125,
+      "completion_length": 543.625,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.3947944939136505,
-      "kl": 0.000660212775983382,
+      "grad_norm": 0.34326109290122986,
+      "kl": 0.0008119463309412822,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 2.5195686295628548,
-      "reward_std": 2.361702173948288,
-      "rewards/concensus_correctness_reward_func": 1.25,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 0.920028967782855,
+      "reward_std": 1.0396093279123306,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.6981310546398163,
-      "rewards/soft_format_reward_func": 0.015625,
-      "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": -0.03793750610202551,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6296851858496666,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.024718777276575565,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 7.443942479312682e-07,
-      "train_runtime": 611.3719,
-      "train_samples_per_second": 0.523,
-      "train_steps_per_second": 0.033
+      "train_loss": 7.208689851267991e-07,
+      "train_runtime": 624.2234,
+      "train_samples_per_second": 0.513,
+      "train_steps_per_second": 0.032
     }
   ],
   "logging_steps": 2,