End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e1728362b9231b39a86fabbcf8b3429f649f7197724fb04438c2b8f86f576410
+oid sha256:9cb947b986d4af1f0241989f8bad8dced2ee738fc82df8633f1d960f0f9f0af1
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.7891535435410333e-06,
-    "train_runtime": 610.1098,
+    "train_loss": 1.5873314964665042e-06,
+    "train_runtime": 578.0843,
     "train_samples": 11,
-    "train_samples_per_second": 0.524,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.554,
+    "train_steps_per_second": 0.035
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.7891535435410333e-06,
-    "train_runtime": 610.1098,
+    "train_loss": 1.5873314964665042e-06,
+    "train_runtime": 578.0843,
     "train_samples": 11,
-    "train_samples_per_second": 0.524,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.554,
+    "train_steps_per_second": 0.035
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 432.3125,
+      "completion_length": 489.59375,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.41115620732307434,
-      "kl": 0.001654699444770813,
+      "grad_norm": 1.5638442039489746,
+      "kl": 0.0014776699608773924,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 1.0831282511353493,
-      "reward_std": 1.0800125412642956,
-      "rewards/concensus_correctness_reward_func": 0.06375000067055225,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 0.8261259887367487,
+      "reward_std": 1.043325025588274,
+      "rewards/concensus_correctness_reward_func": 0.10949999839067459,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.493065744638443,
+      "rewards/question_recreation_reward_func": 0.5658447369933128,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.08881250536069274,
+      "rewards/xmlcount_reward_func": -0.03671875223517418,
       "step": 2
     },
     {
-      "completion_length": 495.32142857142856,
+      "completion_length": 507.2142857142857,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.4448104798793793,
-      "kl": 0.001431963880479868,
+      "grad_norm": 0.38710102438926697,
+      "kl": 0.0014355787674763373,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.7858572985444751,
-      "reward_std": 0.8992064595222473,
-      "rewards/concensus_correctness_reward_func": 0.04171428403684071,
+      "reward": 0.6985892738614764,
+      "reward_std": 1.0949053594044276,
+      "rewards/concensus_correctness_reward_func": 0.006499999868018287,
       "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5004644564219883,
+      "rewards/question_recreation_reward_func": 0.5299107347215924,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.17224999623639242,
+      "rewards/xmlcount_reward_func": 0.09075000243527549,
       "step": 4
     },
     {
-      "completion_length": 427.42857142857144,
+      "completion_length": 464.7857142857143,
       "epoch": 2.0,
-      "grad_norm": 0.35848864912986755,
-      "kl": 0.0016491883101740054,
+      "grad_norm": 0.38429608941078186,
+      "kl": 0.0013789832924625703,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.1533292021070207,
-      "reward_std": 0.9785437073026385,
-      "rewards/concensus_correctness_reward_func": 0.00964285752602986,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 1.213138461112976,
+      "reward_std": 0.9307111118521009,
+      "rewards/concensus_correctness_reward_func": 0.1446428554398673,
+      "rewards/consensus_reward_func": 0.5714285714285714,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5021148749760219,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.48256705488477436,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2130000085702964,
+      "rewards/xmlcount_reward_func": 0.014500010226454054,
       "step": 6
     },
     {
-      "completion_length": 396.03125,
+      "completion_length": 337.5625,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.4652974009513855,
-      "kl": 0.0018128370720660314,
+      "grad_norm": 0.5432289838790894,
+      "kl": 0.001892338739708066,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.1249517127871513,
-      "reward_std": 0.972180338576436,
-      "rewards/concensus_correctness_reward_func": 0.0426874989643693,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 0.8827108982950449,
+      "reward_std": 0.8278120644390583,
+      "rewards/concensus_correctness_reward_func": 0.0651249997317791,
+      "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5647642109543085,
+      "rewards/question_recreation_reward_func": 0.4585858900099993,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.2206249968148768,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.21837500296533108,
       "step": 8
     },
     {
-      "completion_length": 507.0357142857143,
+      "completion_length": 468.5357142857143,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.4943205416202545,
-      "kl": 0.0017367783168862974,
+      "grad_norm": 0.405779093503952,
+      "kl": 0.001528177410364151,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.1879901460238866,
-      "reward_std": 1.009871048586709,
-      "rewards/concensus_correctness_reward_func": 0.07971428547586713,
+      "reward": 0.9205435940197536,
+      "reward_std": 1.0938205633844649,
+      "rewards/concensus_correctness_reward_func": 0.006499999868018287,
       "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5498829739434379,
+      "rewards/question_recreation_reward_func": 0.4201864632112639,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.30839285786662785,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.26171428178037914,
       "step": 10
     },
     {
-      "completion_length": 424.85714285714283,
+      "completion_length": 435.0,
       "epoch": 4.0,
-      "grad_norm": 0.541883111000061,
-      "kl": 0.0019839617057836483,
+      "grad_norm": 0.38847753405570984,
+      "kl": 0.0015181111770548991,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.0200607010296412,
-      "reward_std": 0.9637793579271862,
-      "rewards/concensus_correctness_reward_func": 0.0825000022138868,
-      "rewards/consensus_reward_func": 0.42857142857142855,
+      "reward": 0.9678261620657784,
+      "reward_std": 0.7374571923698697,
+      "rewards/concensus_correctness_reward_func": 0.06257142765181405,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4172749881233488,
+      "rewards/question_recreation_reward_func": 0.4248975728239332,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.09171428637845176,
+      "rewards/xmlcount_reward_func": 0.4089285837752478,
       "step": 12
     },
     {
-      "completion_length": 405.125,
+      "completion_length": 419.34375,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.7639334797859192,
-      "kl": 0.003259157674619928,
+      "grad_norm": 0.45804429054260254,
+      "kl": 0.0018354781859670766,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.3091236352920532,
-      "reward_std": 1.4509413577616215,
-      "rewards/concensus_correctness_reward_func": 0.1508749984204769,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 1.1636791788041592,
+      "reward_std": 1.2268624641001225,
+      "rewards/concensus_correctness_reward_func": 0.021312499884516,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.44602989312261343,
-      "rewards/soft_format_reward_func": 0.015625,
-      "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": 0.2903437493368983,
+      "rewards/question_recreation_reward_func": 0.5328666642308235,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.2813750021159649,
       "step": 14
     },
     {
-      "completion_length": 403.42857142857144,
+      "completion_length": 382.07142857142856,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.4199073314666748,
-      "kl": 0.0016617131186649203,
+      "grad_norm": 0.4776652157306671,
+      "kl": 0.0018598723358341626,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.9939025385039193,
-      "reward_std": 1.1395008031811034,
-      "rewards/concensus_correctness_reward_func": 0.05021428742579052,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 1.2463292011192866,
+      "reward_std": 1.1558310389518738,
+      "rewards/concensus_correctness_reward_func": 0.019500000136239187,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.3302239828876087,
+      "rewards/question_recreation_reward_func": 0.40829351437943323,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.23846428734915598,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.3185357249208859,
       "step": 16
     },
     {
-      "completion_length": 414.17857142857144,
+      "completion_length": 486.67857142857144,
       "epoch": 6.0,
-      "grad_norm": 0.9216535091400146,
-      "kl": 0.0019826509690444383,
+      "grad_norm": 21.22762107849121,
+      "kl": 0.002586001729858773,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 0.7129433985267367,
-      "reward_std": 0.8727979681321553,
-      "rewards/concensus_correctness_reward_func": 0.01928571505205972,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 1.0712240423474992,
+      "reward_std": 0.6037323964493615,
+      "rewards/concensus_correctness_reward_func": 0.0346428571002824,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.37819340346114977,
+      "rewards/question_recreation_reward_func": 0.5184025785752705,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.0297500022820064,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.42889287216322763,
       "step": 18
     },
     {
-      "completion_length": 460.0,
+      "completion_length": 514.15625,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.48125284910202026,
-      "kl": 0.001955981715582311,
+      "grad_norm": 0.3629147708415985,
+      "kl": 0.0016208993110922165,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.39322067797184,
-      "reward_std": 1.2507520020008087,
-      "rewards/concensus_correctness_reward_func": 0.08237500116229057,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 1.2288756389170885,
+      "reward_std": 1.051327284425497,
+      "rewards/concensus_correctness_reward_func": 0.0296249995008111,
+      "rewards/consensus_reward_func": 0.4375,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5450331903994083,
+      "rewards/question_recreation_reward_func": 0.6005943901836872,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.26581248827278614,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.12990625528618693,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 1.7891535435410333e-06,
-      "train_runtime": 610.1098,
-      "train_samples_per_second": 0.524,
-      "train_steps_per_second": 0.033
+      "train_loss": 1.5873314964665042e-06,
+      "train_runtime": 578.0843,
+      "train_samples_per_second": 0.554,
+      "train_steps_per_second": 0.035
     }
   ],
   "logging_steps": 2,