End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:917718df5314f9dcc64d86a654c1666dc0e0bc9667c55f8ed9f5157383594f77
+oid sha256:98b773de19e714ae063411e69dd63c3b2da5fc7649d1da5a6ac967c2439c727e
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.3379344068198407e-06,
-    "train_runtime": 591.9354,
+    "train_loss": 1.4172433793646632e-06,
+    "train_runtime": 680.0631,
     "train_samples": 11,
-    "train_samples_per_second": 0.541,
-    "train_steps_per_second": 0.034
+    "train_samples_per_second": 0.471,
+    "train_steps_per_second": 0.029
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.3379344068198407e-06,
-    "train_runtime": 591.9354,
+    "train_loss": 1.4172433793646632e-06,
+    "train_runtime": 680.0631,
     "train_samples": 11,
-    "train_samples_per_second": 0.541,
-    "train_steps_per_second": 0.034
+    "train_samples_per_second": 0.471,
+    "train_steps_per_second": 0.029
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 550.71875,
+      "completion_length": 472.3125,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.5056862831115723,
-      "kl": 0.0013800587548757903,
+      "grad_norm": 1.2652982473373413,
+      "kl": 0.001695555263722781,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 1.312247209250927,
-      "reward_std": 1.1249291487038136,
-      "rewards/concensus_correctness_reward_func": 0.026562499348074198,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 0.781109157949686,
+      "reward_std": 1.008236899971962,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6732159778475761,
+      "rewards/question_recreation_reward_func": 0.46079665049910545,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.22184374369680882,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.22656250279396772,
       "step": 2
     },
     {
-      "completion_length": 415.64285714285717,
+      "completion_length": 583.2142857142857,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.4648090600967407,
-      "kl": 0.001704983363327171,
+      "grad_norm": 0.5107342004776001,
+      "kl": 0.00136417372518086,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 1.271498522588185,
-      "reward_std": 1.0002022215298243,
-      "rewards/concensus_correctness_reward_func": 0.07014285719820432,
-      "rewards/consensus_reward_func": 0.42857142857142855,
+      "reward": 0.49905759734766825,
+      "reward_std": 0.9771258064678737,
+      "rewards/concensus_correctness_reward_func": 0.03157142869063786,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5835699651922498,
+      "rewards/question_recreation_reward_func": 0.5851290098258427,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.189214289188385,
+      "rewards/xmlcount_reward_func": -0.33192857035568785,
       "step": 4
     },
     {
-      "completion_length": 378.5,
+      "completion_length": 456.2857142857143,
       "epoch": 2.0,
-      "grad_norm": 0.6165847182273865,
-      "kl": 0.001546859059349767,
+      "grad_norm": 0.46626871824264526,
+      "kl": 0.0014885144274947898,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.2759263174874442,
-      "reward_std": 0.8719319019998822,
-      "rewards/concensus_correctness_reward_func": 0.0247142847095217,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 0.9597211565290179,
+      "reward_std": 1.0505088099411555,
+      "rewards/concensus_correctness_reward_func": 0.03157142869063786,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5256763185773577,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5903996995517186,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.05357142857142857,
-      "rewards/xmlcount_reward_func": 0.314821436469044,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.1591785635266985,
       "step": 6
     },
     {
-      "completion_length": 529.46875,
+      "completion_length": 534.875,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.28671154379844666,
-      "kl": 0.001089718221919611,
+      "grad_norm": 0.5832344889640808,
+      "kl": 0.0015865868335822597,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.3038343638181686,
-      "reward_std": 1.325776332989335,
-      "rewards/concensus_correctness_reward_func": 0.11287499964237213,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 0.7514080386608839,
+      "reward_std": 0.9840278550982475,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.6289281100034714,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5922518000006676,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.00046874722465872765,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.04978124983608723,
       "step": 8
     },
     {
-      "completion_length": 417.35714285714283,
+      "completion_length": 616.0357142857143,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.568111002445221,
-      "kl": 0.001521418145525136,
+      "grad_norm": 0.29018133878707886,
+      "kl": 0.0012787416593969933,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.1384957901069097,
-      "reward_std": 1.1431427342551095,
-      "rewards/concensus_correctness_reward_func": 0.06449999979564122,
+      "reward": 0.8998302583183561,
+      "reward_std": 0.892537602356502,
+      "rewards/concensus_correctness_reward_func": 0.09642857100282397,
       "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.42867439133780344,
+      "rewards/question_recreation_reward_func": 0.5202945108924594,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.07142857142857142,
-      "rewards/xmlcount_reward_func": 0.43103571449007305,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.1402500026992389,
       "step": 10
     },
     {
-      "completion_length": 510.25,
+      "completion_length": 465.7142857142857,
       "epoch": 4.0,
-      "grad_norm": 0.34620562195777893,
-      "kl": 0.0014288202510215342,
+      "grad_norm": 0.3647068738937378,
+      "kl": 0.0018004042794927955,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.0430249444075994,
-      "reward_std": 1.2671920103686196,
-      "rewards/concensus_correctness_reward_func": 0.005642857402563095,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 1.0628357018743242,
+      "reward_std": 1.1306542732885905,
+      "rewards/concensus_correctness_reward_func": 0.01578571434531893,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5668106632573264,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5234428378088134,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.04200000422341483,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.2736071305615561,
       "step": 12
     },
     {
-      "completion_length": 515.09375,
+      "completion_length": 559.25,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.370266854763031,
-      "kl": 0.0012138541715103202,
+      "grad_norm": 0.40000659227371216,
+      "kl": 0.0013819929226883687,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.2918198108673096,
-      "reward_std": 1.2725212015211582,
-      "rewards/concensus_correctness_reward_func": 0.004937500227242708,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.978723555803299,
+      "reward_std": 1.2774476408958435,
+      "rewards/concensus_correctness_reward_func": 0.05599999986588955,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.6141322962939739,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5877548195421696,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.28212498873472214,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.08496874943375587,
       "step": 14
     },
     {
-      "completion_length": 426.9642857142857,
+      "completion_length": 591.4285714285714,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.5171462297439575,
-      "kl": 0.0014096008796643997,
+      "grad_norm": 0.33472514152526855,
+      "kl": 0.0013038381418612386,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.907790060554232,
-      "reward_std": 0.8804280289581844,
-      "rewards/concensus_correctness_reward_func": 0.0,
+      "reward": 0.938178539276123,
+      "reward_std": 1.1371430669512068,
+      "rewards/concensus_correctness_reward_func": 0.03157142869063786,
       "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5585757579122271,
-      "rewards/soft_format_reward_func": 0.017857142857142856,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.11707142421177455,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5574999366487775,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.1705357219491686,
       "step": 16
     },
     {
-      "completion_length": 478.5,
+      "completion_length": 452.10714285714283,
       "epoch": 6.0,
-      "grad_norm": 3.2752087116241455,
-      "kl": 0.001824594086169132,
+      "grad_norm": 0.40762996673583984,
+      "kl": 0.001767604670021683,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.1057345696857996,
-      "reward_std": 0.8967528258051191,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 0.9837531319686345,
+      "reward_std": 1.1079531908035278,
+      "rewards/concensus_correctness_reward_func": 0.01578571434531893,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.46877026770796093,
+      "rewards/question_recreation_reward_func": 0.4228245658533914,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.2798214320625578,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.2951428560273988,
       "step": 18
     },
     {
-      "completion_length": 383.78125,
+      "completion_length": 534.5,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.4814499020576477,
-      "kl": 0.0014182818413246423,
+      "grad_norm": 0.38121506571769714,
+      "kl": 0.0015870211573201232,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.2847145721316338,
-      "reward_std": 0.9296081736683846,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 1.057984434068203,
+      "reward_std": 1.0102076008915901,
+      "rewards/concensus_correctness_reward_func": 0.05525000020861626,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.47596458345651627,
+      "rewards/question_recreation_reward_func": 0.61589065939188,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.3556249951943755,
+      "rewards/xmlcount_reward_func": 0.12121874652802944,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 1.3379344068198407e-06,
-      "train_runtime": 591.9354,
-      "train_samples_per_second": 0.541,
-      "train_steps_per_second": 0.034
+      "train_loss": 1.4172433793646632e-06,
+      "train_runtime": 680.0631,
+      "train_samples_per_second": 0.471,
+      "train_steps_per_second": 0.029
     }
   ],
   "logging_steps": 2,