End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:98cfde7df46d4c7e48dd753ebe49b8f2ee98d0d9239c65df96bf37f505ab4bde
+oid sha256:fe97180567c108c8b034cd043c2b90fbf8cf865313619a5078a33faae922a57d
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 9.91116530713043e-07,
-    "train_runtime": 599.1444,
+    "train_loss": 1.3643185923228884e-06,
+    "train_runtime": 527.5941,
     "train_samples": 11,
-    "train_samples_per_second": 0.534,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.607,
+    "train_steps_per_second": 0.038
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 9.91116530713043e-07,
-    "train_runtime": 599.1444,
+    "train_loss": 1.3643185923228884e-06,
+    "train_runtime": 527.5941,
     "train_samples": 11,
-    "train_samples_per_second": 0.534,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.607,
+    "train_steps_per_second": 0.038
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 412.71875,
+      "completion_length": 412.625,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.6986134052276611,
-      "kl": 0.000978747797489632,
+      "grad_norm": 0.43695393204689026,
+      "kl": 0.0009111311810556799,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 3.4132684394717216,
-      "reward_std": 2.4529526382684708,
-      "rewards/concensus_correctness_reward_func": 1.9297499991953373,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 1.229908674955368,
+      "reward_std": 0.8567641489207745,
+      "rewards/concensus_correctness_reward_func": 0.029687500558793545,
+      "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6463933773338795,
+      "rewards/question_recreation_reward_func": 0.6100648939609528,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.3371250033378601,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.4495312590152025,
       "step": 2
     },
     {
-      "completion_length": 475.35714285714283,
+      "completion_length": 372.67857142857144,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.6639646291732788,
-      "kl": 0.0008992253736193691,
+      "grad_norm": 0.3806152045726776,
+      "kl": 0.0010995941286507463,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.725197604724339,
-      "reward_std": 1.121276923588344,
-      "rewards/concensus_correctness_reward_func": -0.09235714163099017,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 0.8413883000612259,
+      "reward_std": 0.7054694252354758,
+      "rewards/concensus_correctness_reward_func": 0.044857144355773926,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.3751618968588965,
+      "rewards/question_recreation_reward_func": 0.44474541928086964,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.08525000725473676,
+      "rewards/xmlcount_reward_func": 0.13750000623986125,
       "step": 4
     },
     {
-      "completion_length": 388.5357142857143,
+      "completion_length": 314.5,
       "epoch": 2.0,
-      "grad_norm": 0.7052159905433655,
-      "kl": 0.0010295261662187321,
+      "grad_norm": 0.39608466625213623,
+      "kl": 0.0008572574172701154,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 2.9794775588171825,
-      "reward_std": 3.073520983968462,
-      "rewards/concensus_correctness_reward_func": 1.4911428562232427,
-      "rewards/consensus_reward_func": 0.5714285714285714,
+      "reward": 0.9021432910646711,
+      "reward_std": 0.7128424601895469,
+      "rewards/concensus_correctness_reward_func": 0.015285714396408625,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5606560153620583,
+      "rewards/question_recreation_reward_func": 0.5191789876137461,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.3562500221388681,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.2069643001471247,
       "step": 6
     },
     {
-      "completion_length": 409.15625,
+      "completion_length": 455.75,
       "epoch": 2.7272727272727275,
-      "grad_norm": 2.0488579273223877,
-      "kl": 0.0009390798149979673,
+      "grad_norm": 0.45465153455734253,
+      "kl": 0.0008965662418631837,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 2.5797567814588547,
-      "reward_std": 2.7606934532523155,
-      "rewards/concensus_correctness_reward_func": 1.2508125000167638,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 1.0771191604435444,
+      "reward_std": 0.9723493047058582,
+      "rewards/concensus_correctness_reward_func": 0.07962500024586916,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5778506472706795,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6189628974534571,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.18859375827014446,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.11290625296533108,
       "step": 8
     },
     {
-      "completion_length": 413.85714285714283,
+      "completion_length": 346.4642857142857,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.7300273180007935,
-      "kl": 0.0009499788284301758,
+      "grad_norm": 0.4408261477947235,
+      "kl": 0.0010133750038221478,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 0.9401203138487679,
-      "reward_std": 1.399593813078744,
-      "rewards/concensus_correctness_reward_func": 0.1009285684142794,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 1.3129970261028834,
+      "reward_std": 0.863572780575071,
+      "rewards/concensus_correctness_reward_func": 0.026714285569531575,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.3889774701425007,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.4315684565476009,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.07521429232188634,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.49757144280842375,
       "step": 10
     },
     {
-      "completion_length": 442.5357142857143,
+      "completion_length": 449.14285714285717,
       "epoch": 4.0,
-      "grad_norm": 10.092423439025879,
-      "kl": 0.0013360294147527643,
+      "grad_norm": 0.338410347700119,
+      "kl": 0.0008665725638690804,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 3.450189105101994,
-      "reward_std": 2.2575139233044217,
-      "rewards/concensus_correctness_reward_func": 2.0822142865508795,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 1.2780203989573888,
+      "reward_std": 0.7916043145315987,
+      "rewards/concensus_correctness_reward_func": 0.028714286429541453,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5920107013412884,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5381632383380618,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.20453571315322602,
+      "rewards/xmlcount_reward_func": 0.5682857079165322,
       "step": 12
     },
     {
-      "completion_length": 488.03125,
+      "completion_length": 440.15625,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.5048688650131226,
-      "kl": 0.001014460351143498,
+      "grad_norm": 0.5608338713645935,
+      "kl": 0.0009716439744806848,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 0.37938143499195576,
-      "reward_std": 0.7001448348164558,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0,
+      "reward": 1.0658384785056114,
+      "reward_std": 1.0223564468324184,
+      "rewards/concensus_correctness_reward_func": 0.07437500078231096,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.44478767924010754,
+      "rewards/question_recreation_reward_func": 0.5121509656310081,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": -0.08103126846253872,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.26056250277906656,
       "step": 14
     },
     {
-      "completion_length": 588.0714285714286,
+      "completion_length": 418.2857142857143,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.5584092140197754,
-      "kl": 0.0017077010873306011,
+      "grad_norm": 29.862661361694336,
+      "kl": 0.006334888116855707,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 3.4658961040633067,
-      "reward_std": 2.493277262364115,
-      "rewards/concensus_correctness_reward_func": 2.1120714289801463,
-      "rewards/consensus_reward_func": 0.42857142857142855,
+      "reward": 0.997036657163075,
+      "reward_std": 1.0026466037545885,
+      "rewards/concensus_correctness_reward_func": 0.024071428392614638,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.6585390312331063,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5370009754385267,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.12385716182844979,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.2038214217339243,
       "step": 16
     },
     {
-      "completion_length": 444.2857142857143,
+      "completion_length": 400.0,
       "epoch": 6.0,
-      "grad_norm": 0.3846805691719055,
-      "kl": 0.0008876290937353458,
+      "grad_norm": 0.31728827953338623,
+      "kl": 0.001013818895444274,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 2.695590606757573,
-      "reward_std": 2.672016684498106,
-      "rewards/concensus_correctness_reward_func": 1.3670000008174352,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 1.0134610363415308,
+      "reward_std": 0.8706743163721902,
+      "rewards/concensus_correctness_reward_func": 0.03971428530556815,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5523763245769909,
+      "rewards/question_recreation_reward_func": 0.5667110521878514,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2762142781700407,
+      "rewards/xmlcount_reward_func": 0.3356071412563324,
       "step": 18
     },
     {
-      "completion_length": 480.125,
+      "completion_length": 425.75,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.4312440752983093,
-      "kl": 0.0009752309269970283,
+      "grad_norm": 0.4884311854839325,
+      "kl": 0.001051145576639101,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.1490149404853582,
-      "reward_std": 1.1749461703002453,
-      "rewards/concensus_correctness_reward_func": 0.08912499924190342,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 0.91310590878129,
+      "reward_std": 0.8083030693233013,
+      "rewards/concensus_correctness_reward_func": 0.00774999987334013,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.514139941893518,
+      "rewards/question_recreation_reward_func": 0.445574632845819,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.10824999678879976,
+      "rewards/xmlcount_reward_func": 0.3972812704741955,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 9.91116530713043e-07,
-      "train_runtime": 599.1444,
-      "train_samples_per_second": 0.534,
-      "train_steps_per_second": 0.033
+      "train_loss": 1.3643185923228884e-06,
+      "train_runtime": 527.5941,
+      "train_samples_per_second": 0.607,
+      "train_steps_per_second": 0.038
     }
   ],
   "logging_steps": 2,