End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e929e21f855e5010a72bfd5e6df06560a3b9b5eef363fb63697a1a8e818bf91f
+oid sha256:f8c428c4e0594e8b6419bc3d1aee70961a86b8a6ea4e95a7d5615683e87c9199
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.247190754038456e-07,
-    "train_runtime": 693.0986,
+    "train_loss": 6.668610041060674e-07,
+    "train_runtime": 687.7408,
     "train_samples": 11,
-    "train_samples_per_second": 0.462,
+    "train_samples_per_second": 0.465,
     "train_steps_per_second": 0.029
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.247190754038456e-07,
-    "train_runtime": 693.0986,
+    "train_loss": 6.668610041060674e-07,
+    "train_runtime": 687.7408,
     "train_samples": 11,
-    "train_samples_per_second": 0.462,
+    "train_samples_per_second": 0.465,
     "train_steps_per_second": 0.029
 }

trainer_state.json

@@ -10,202 +10,202 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 600.5625,
+      "completion_length": 482.625,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.28632819652557373,
-      "kl": 0.0006985599466133863,
+      "grad_norm": 0.37514087557792664,
+      "kl": 0.0006302265792328399,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 1.7123659923672676,
-      "reward_std": 2.23521476238966,
-      "rewards/concensus_correctness_reward_func": 0.628749999916181,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.8577204570174217,
+      "reward_std": 1.0115732066333294,
+      "rewards/concensus_correctness_reward_func": 0.027000000700354576,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5244909711182117,
+      "rewards/question_recreation_reward_func": 0.5929392203688622,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.18412499874830246,
+      "rewards/xmlcount_reward_func": -0.012218741700053215,
       "step": 2
     },
     {
-      "completion_length": 604.0357142857143,
+      "completion_length": 601.2142857142857,
       "epoch": 1.3636363636363638,
-      "grad_norm": 1.0203968286514282,
-      "kl": 0.0006703816720151476,
+      "grad_norm": 0.39804214239120483,
+      "kl": 0.0006456053649474468,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 1.510402568749019,
-      "reward_std": 2.4836114091532573,
-      "rewards/concensus_correctness_reward_func": 0.7142857142857143,
+      "reward": 0.9125474521092006,
+      "reward_std": 1.0302738632474626,
+      "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": 0.5013311292443957,
+      "rewards/question_recreation_reward_func": 0.45147604175976347,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.04478571563959122,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.24678572160857065,
       "step": 4
     },
     {
-      "completion_length": 605.9642857142857,
+      "completion_length": 508.75,
       "epoch": 2.0,
-      "grad_norm": 0.22705680131912231,
-      "kl": 0.0006028824552361454,
+      "grad_norm": 2.6987972259521484,
+      "kl": 0.0006489827480566289,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.001860580274037,
-      "reward_std": 0.9050898615803037,
-      "rewards/concensus_correctness_reward_func": 0.04785714378314359,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 1.230867428439004,
+      "reward_std": 0.9484705371516091,
+      "rewards/concensus_correctness_reward_func": 0.06671428680419922,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6388605747904096,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.6639745959213802,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.029428571462631226,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.3394642791577748,
       "step": 6
     },
     {
-      "completion_length": 555.84375,
+      "completion_length": 431.84375,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.39249709248542786,
-      "kl": 0.0005997194311930798,
+      "grad_norm": 0.45506787300109863,
+      "kl": 0.0006626314789173193,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 2.719204457476735,
-      "reward_std": 2.539236516226083,
-      "rewards/concensus_correctness_reward_func": 1.2643750000279397,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 0.8808938264846802,
+      "reward_std": 0.7278284281492233,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.6168294418603182,
-      "rewards/soft_format_reward_func": 0.015625,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.30675000324845314,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5341125726699829,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.15928125381469727,
       "step": 8
     },
     {
-      "completion_length": 585.2857142857143,
+      "completion_length": 465.57142857142856,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.6291311383247375,
-      "kl": 0.0006507792110954012,
+      "grad_norm": 8.681121826171875,
+      "kl": 0.001499615642907364,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 0.9478132937635694,
-      "reward_std": 1.137954294681549,
-      "rewards/concensus_correctness_reward_func": 0.007142856983201844,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 1.159986334187644,
+      "reward_std": 0.9050823748111725,
+      "rewards/concensus_correctness_reward_func": 0.030857143657548085,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6311704346111843,
+      "rewards/question_recreation_reward_func": 0.5218434844698224,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.005928567477634975,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.42871427855321337,
       "step": 10
     },
     {
-      "completion_length": 545.0,
+      "completion_length": 509.89285714285717,
       "epoch": 4.0,
-      "grad_norm": 0.5774611830711365,
-      "kl": 0.0007185281470551022,
+      "grad_norm": 0.29260939359664917,
+      "kl": 0.0006353700092794108,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 0.6356133509959493,
-      "reward_std": 1.2423699987786156,
-      "rewards/concensus_correctness_reward_func": 0.08714285918644496,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 0.8939749896526337,
+      "reward_std": 0.841909938624927,
+      "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": 0.5161490387150219,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6119035524981362,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.39625003029193195,
+      "rewards/xmlcount_reward_func": 0.2106428529535021,
       "step": 12
     },
     {
-      "completion_length": 531.375,
+      "completion_length": 486.96875,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.4670509099960327,
-      "kl": 0.0007279670171556063,
+      "grad_norm": 0.34878578782081604,
+      "kl": 0.0006228954916878138,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.021582493558526,
-      "reward_std": 0.982480026781559,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0,
+      "reward": 1.3040558397769928,
+      "reward_std": 1.0455535128712654,
+      "rewards/concensus_correctness_reward_func": 0.026499999687075615,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.6383012235164642,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6183371227234602,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.24265624955296516,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.4092187471687794,
       "step": 14
     },
     {
-      "completion_length": 564.4285714285714,
+      "completion_length": 538.3928571428571,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.5724964141845703,
-      "kl": 0.0006105650058348797,
+      "grad_norm": 0.3759525418281555,
+      "kl": 0.0006132706891678806,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 1.5704969838261604,
-      "reward_std": 2.1429328663008556,
-      "rewards/concensus_correctness_reward_func": 0.7142857142857143,
+      "reward": 1.2082966821534293,
+      "reward_std": 0.5831507540174893,
+      "rewards/concensus_correctness_reward_func": 0.030285713928086416,
       "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5301398145301002,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6400823678289141,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.022500004087175642,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.3236428403428623,
       "step": 16
     },
     {
-      "completion_length": 567.75,
+      "completion_length": 679.1071428571429,
       "epoch": 6.0,
-      "grad_norm": 0.2651831805706024,
-      "kl": 0.0007271821211491312,
+      "grad_norm": 0.3067660629749298,
+      "kl": 0.0005424516608140298,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 0.638118999344962,
-      "reward_std": 1.099077263048717,
+      "reward": 0.7338065419878278,
+      "reward_std": 0.9008073381015232,
       "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": 0.5106904400246484,
+      "rewards/question_recreation_reward_func": 0.6232708266803196,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.05357142857142857,
-      "rewards/xmlcount_reward_func": 0.07385714777878352,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.11053570359945297,
       "step": 18
     },
     {
-      "completion_length": 517.5625,
+      "completion_length": 550.125,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.42876189947128296,
-      "kl": 0.0006498598304460756,
+      "grad_norm": 0.3706094026565552,
+      "kl": 0.0006239530994207598,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 2.6714578419923782,
-      "reward_std": 2.602976344525814,
-      "rewards/concensus_correctness_reward_func": 1.253749999916181,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 1.1454841140657663,
+      "reward_std": 0.9183149933815002,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.3125,
-      "rewards/question_recreation_reward_func": 0.5567391663789749,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5890153739601374,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2984687499701977,
+      "rewards/xmlcount_reward_func": 0.2439687605947256,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 6.247190754038456e-07,
-      "train_runtime": 693.0986,
-      "train_samples_per_second": 0.462,
+      "train_loss": 6.668610041060674e-07,
+      "train_runtime": 687.7408,
+      "train_samples_per_second": 0.465,
       "train_steps_per_second": 0.029
     }
   ],