End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8c428c4e0594e8b6419bc3d1aee70961a86b8a6ea4e95a7d5615683e87c9199
+oid sha256:cd98e00b0eb5f1b79ef312cd4553a048ec68874e8fb6ac1f417df77ac2a102c8
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.668610041060674e-07,
-    "train_runtime": 687.7408,
+    "train_loss": 6.243158054530796e-07,
+    "train_runtime": 662.6617,
     "train_samples": 11,
-    "train_samples_per_second": 0.465,
-    "train_steps_per_second": 0.029
+    "train_samples_per_second": 0.483,
+    "train_steps_per_second": 0.03
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.668610041060674e-07,
-    "train_runtime": 687.7408,
+    "train_loss": 6.243158054530796e-07,
+    "train_runtime": 662.6617,
     "train_samples": 11,
-    "train_samples_per_second": 0.465,
-    "train_steps_per_second": 0.029
+    "train_samples_per_second": 0.483,
+    "train_steps_per_second": 0.03
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 482.625,
+      "completion_length": 518.59375,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.37514087557792664,
-      "kl": 0.0006302265792328399,
+      "grad_norm": 0.3101458251476288,
+      "kl": 0.0006586692034034058,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 0.8577204570174217,
-      "reward_std": 1.0115732066333294,
-      "rewards/concensus_correctness_reward_func": 0.027000000700354576,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 1.5320980101823807,
+      "reward_std": 1.391620872542262,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5929392203688622,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5687229856848717,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.012218741700053215,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.4321250058710575,
       "step": 2
     },
     {
-      "completion_length": 601.2142857142857,
+      "completion_length": 607.6785714285714,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.39804214239120483,
-      "kl": 0.0006456053649474468,
+      "grad_norm": 0.41752761602401733,
+      "kl": 0.0005801213389661695,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.9125474521092006,
-      "reward_std": 1.0302738632474626,
+      "reward": 0.7936271301337651,
+      "reward_std": 0.8227866802896772,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.45147604175976347,
+      "rewards/question_recreation_reward_func": 0.5946271462099892,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.24678572160857065,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.03828570778880801,
       "step": 4
     },
     {
-      "completion_length": 508.75,
+      "completion_length": 419.9642857142857,
       "epoch": 2.0,
-      "grad_norm": 2.6987972259521484,
-      "kl": 0.0006489827480566289,
+      "grad_norm": 0.6341163516044617,
+      "kl": 0.0007463138733458306,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.230867428439004,
-      "reward_std": 0.9484705371516091,
-      "rewards/concensus_correctness_reward_func": 0.06671428680419922,
-      "rewards/consensus_reward_func": 0.07142857142857142,
+      "reward": 1.005787574819156,
+      "reward_std": 0.9503834183726992,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.6639745959213802,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6678590263639178,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.3394642791577748,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": -0.01921428952898298,
       "step": 6
     },
     {
-      "completion_length": 431.84375,
+      "completion_length": 458.625,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.45506787300109863,
-      "kl": 0.0006626314789173193,
+      "grad_norm": 0.337610125541687,
+      "kl": 0.0006834228988736868,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 0.8808938264846802,
-      "reward_std": 0.7278284281492233,
+      "reward": 1.112733330577612,
+      "reward_std": 0.8835000544786453,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5341125726699829,
+      "rewards/question_recreation_reward_func": 0.6689833588898182,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.15928125381469727,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.24062499403953552,
       "step": 8
     },
     {
-      "completion_length": 465.57142857142856,
+      "completion_length": 511.0357142857143,
       "epoch": 3.3636363636363638,
-      "grad_norm": 8.681121826171875,
-      "kl": 0.001499615642907364,
+      "grad_norm": 0.4655108153820038,
+      "kl": 0.0006738068186677992,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.159986334187644,
-      "reward_std": 0.9050823748111725,
-      "rewards/concensus_correctness_reward_func": 0.030857143657548085,
+      "reward": 0.7448970739330564,
+      "reward_std": 0.9845571688243321,
+      "rewards/concensus_correctness_reward_func": 0.04607142720903669,
       "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5218434844698224,
+      "rewards/question_recreation_reward_func": 0.5173256482396807,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.42871427855321337,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.038642864142145426,
       "step": 10
     },
     {
-      "completion_length": 509.89285714285717,
+      "completion_length": 581.4642857142857,
       "epoch": 4.0,
-      "grad_norm": 0.29260939359664917,
-      "kl": 0.0006353700092794108,
+      "grad_norm": 0.3229527473449707,
+      "kl": 0.0006438559919063534,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 0.8939749896526337,
-      "reward_std": 0.841909938624927,
+      "reward": 0.901876015322549,
+      "reward_std": 0.5733589742864881,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.07142857142857142,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6119035524981362,
+      "rewards/question_recreation_reward_func": 0.5391260130064828,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2106428529535021,
+      "rewards/xmlcount_reward_func": 0.21989286371639796,
       "step": 12
     },
     {
-      "completion_length": 486.96875,
+      "completion_length": 471.3125,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.34878578782081604,
-      "kl": 0.0006228954916878138,
+      "grad_norm": 0.45552772283554077,
+      "kl": 0.0006461980010499246,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.3040558397769928,
-      "reward_std": 1.0455535128712654,
-      "rewards/concensus_correctness_reward_func": 0.026499999687075615,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 1.0484812445938587,
+      "reward_std": 0.9308878369629383,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6183371227234602,
+      "rewards/question_recreation_reward_func": 0.5975437425076962,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4092187471687794,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.24781250103842467,
       "step": 14
     },
     {
-      "completion_length": 538.3928571428571,
+      "completion_length": 495.10714285714283,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.3759525418281555,
-      "kl": 0.0006132706891678806,
+      "grad_norm": 16.87542152404785,
+      "kl": 0.0007952758543459433,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 1.2082966821534293,
-      "reward_std": 0.5831507540174893,
-      "rewards/concensus_correctness_reward_func": 0.030285713928086416,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 1.2276427149772644,
+      "reward_std": 0.8300601371697017,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6400823678289141,
+      "rewards/question_recreation_reward_func": 0.5341070124081203,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.3236428403428623,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.372107134865863,
       "step": 16
     },
     {
-      "completion_length": 679.1071428571429,
+      "completion_length": 681.3571428571429,
       "epoch": 6.0,
-      "grad_norm": 0.3067660629749298,
-      "kl": 0.0005424516608140298,
+      "grad_norm": 0.4771121144294739,
+      "kl": 0.000520679762952828,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 0.7338065419878278,
-      "reward_std": 0.9008073381015232,
+      "reward": 0.602795483810561,
+      "reward_std": 0.9616657410349164,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_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": 0.6232708266803196,
+      "rewards/question_recreation_reward_func": 0.5815811795847756,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.11053570359945297,
+      "rewards/xmlcount_reward_func": -0.1216428450175694,
       "step": 18
     },
     {
-      "completion_length": 550.125,
+      "completion_length": 476.34375,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.3706094026565552,
-      "kl": 0.0006239530994207598,
+      "grad_norm": 0.4769397974014282,
+      "kl": 0.000715115420462098,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.1454841140657663,
-      "reward_std": 0.9183149933815002,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.5274622952565551,
+      "reward_std": 0.5800921861082315,
+      "rewards/concensus_correctness_reward_func": 0.025499999523162842,
+      "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5890153739601374,
+      "rewards/question_recreation_reward_func": 0.5373685359954834,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2439687605947256,
+      "rewards/xmlcount_reward_func": -0.16040625050663948,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 6.668610041060674e-07,
-      "train_runtime": 687.7408,
-      "train_samples_per_second": 0.465,
-      "train_steps_per_second": 0.029
+      "train_loss": 6.243158054530796e-07,
+      "train_runtime": 662.6617,
+      "train_samples_per_second": 0.483,
+      "train_steps_per_second": 0.03
     }
   ],
   "logging_steps": 2,