End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e11011c8120feceb37592c0c851c70ae396f56d643fe7d5a74c5e8d52a9462f4
+oid sha256:a06e0735fce1268b71e714768fb80177e2a7389b998688a2e3c17e5df5ba5960
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.541969924001023e-07,
-    "train_runtime": 1061.1344,
+    "train_loss": 6.380430704666651e-07,
+    "train_runtime": 1093.548,
     "train_samples": 11,
-    "train_samples_per_second": 0.302,
-    "train_steps_per_second": 0.019
+    "train_samples_per_second": 0.293,
+    "train_steps_per_second": 0.018
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 6.541969924001023e-07,
-    "train_runtime": 1061.1344,
+    "train_loss": 6.380430704666651e-07,
+    "train_runtime": 1093.548,
     "train_samples": 11,
-    "train_samples_per_second": 0.302,
-    "train_steps_per_second": 0.019
+    "train_samples_per_second": 0.293,
+    "train_steps_per_second": 0.018
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 489.75,
+      "completion_length": 529.125,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.473178893327713,
-      "kl": 0.000693090507411398,
+      "grad_norm": 0.4646318852901459,
+      "kl": 0.0006358221507980488,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 1.1078154370188713,
-      "reward_std": 1.2845153734087944,
-      "rewards/concensus_correctness_reward_func": 0.001312499982304871,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.39380765706300735,
+      "reward_std": 0.9743207413703203,
+      "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": 0.5154716987162828,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.4490889124572277,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.20040624728426337,
+      "rewards/xmlcount_reward_func": -0.13340624701231718,
       "step": 2
     },
     {
-      "completion_length": 417.25,
+      "completion_length": 501.5,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.4299594759941101,
-      "kl": 0.0007473819797659027,
+      "grad_norm": 0.4398326277732849,
+      "kl": 0.0006624749346104052,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 4.208853593894413,
-      "reward_std": 2.59207792367254,
-      "rewards/concensus_correctness_reward_func": 2.233499999823315,
-      "rewards/consensus_reward_func": 0.6428571428571429,
+      "reward": 1.6800819635391235,
+      "reward_std": 1.4839196843760354,
+      "rewards/concensus_correctness_reward_func": 0.11485714146069118,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.2857142857142857,
-      "rewards/question_recreation_reward_func": 0.6093176560742515,
+      "rewards/final_correctness_reward_func": 0.21428571428571427,
+      "rewards/question_recreation_reward_func": 0.5619391628674099,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4374642749982221,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.39614285874579636,
       "step": 4
     },
     {
-      "completion_length": 438.67857142857144,
+      "completion_length": 515.0357142857143,
       "epoch": 2.0,
-      "grad_norm": 0.3509567975997925,
-      "kl": 0.0009807755954430572,
+      "grad_norm": 0.3059370517730713,
+      "kl": 0.0007214588612051946,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 2.260260526623045,
-      "reward_std": 2.5936509115355357,
-      "rewards/concensus_correctness_reward_func": 0.7407857137066978,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 0.8787845202854702,
+      "reward_std": 1.4067505257470267,
+      "rewards/concensus_correctness_reward_func": 0.0821428554398673,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.35714285714285715,
-      "rewards/question_recreation_reward_func": 0.3300105184316635,
+      "rewards/final_correctness_reward_func": 0.14285714285714285,
+      "rewards/question_recreation_reward_func": 0.45535593373434885,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.29660714098385405,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": -0.08728571874754769,
       "step": 6
     },
     {
-      "completion_length": 391.84375,
+      "completion_length": 450.46875,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.5801579356193542,
-      "kl": 0.0006373556352627929,
+      "grad_norm": 0.4287567734718323,
+      "kl": 0.0006250313490454573,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 2.149243339896202,
-      "reward_std": 2.3885336834937334,
-      "rewards/concensus_correctness_reward_func": 0.6263124999823049,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 2.1336237341165543,
+      "reward_std": 2.356996573507786,
+      "rewards/concensus_correctness_reward_func": 0.7309999987483025,
+      "rewards/consensus_reward_func": 0.4375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.3125,
-      "rewards/question_recreation_reward_func": 0.526024529710412,
+      "rewards/final_correctness_reward_func": 0.125,
+      "rewards/question_recreation_reward_func": 0.5192175172269344,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.3562812418676913,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.3209062535315752,
       "step": 8
     },
     {
-      "completion_length": 474.7142857142857,
+      "completion_length": 415.60714285714283,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.501632571220398,
-      "kl": 0.0006951373444670546,
+      "grad_norm": 0.4459122121334076,
+      "kl": 0.0009690781423289861,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 3.942079620701926,
-      "reward_std": 2.432533400399344,
-      "rewards/concensus_correctness_reward_func": 2.1785714285714284,
-      "rewards/consensus_reward_func": 0.9285714285714286,
+      "reward": 1.4117277903216225,
+      "reward_std": 1.060990469796317,
+      "rewards/concensus_correctness_reward_func": 0.12857142516544887,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.21428571428571427,
-      "rewards/question_recreation_reward_func": 0.4000081185783659,
+      "rewards/question_recreation_reward_func": 0.5573349041598183,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.20278571333203996,
+      "rewards/xmlcount_reward_func": 0.27939286615167347,
       "step": 10
     },
     {
-      "completion_length": 580.6785714285714,
+      "completion_length": 584.9642857142857,
       "epoch": 4.0,
-      "grad_norm": 0.4899674952030182,
-      "kl": 0.0006590454681177757,
+      "grad_norm": 0.36204561591148376,
+      "kl": 0.000589396118552291,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.1509517090661185,
-      "reward_std": 1.116724627358573,
-      "rewards/concensus_correctness_reward_func": 0.004500000072377068,
-      "rewards/consensus_reward_func": 0.42857142857142855,
+      "reward": 0.9562124311923981,
+      "reward_std": 1.1778249740600586,
+      "rewards/concensus_correctness_reward_func": 0.08471428496497017,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.45245168251650675,
+      "rewards/question_recreation_reward_func": 0.481891040291105,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.26542856437819345,
+      "rewards/xmlcount_reward_func": 0.03246427859578814,
       "step": 12
     },
     {
-      "completion_length": 399.34375,
+      "completion_length": 432.65625,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.5264378786087036,
-      "kl": 0.0007004420804150868,
+      "grad_norm": 0.5002179741859436,
+      "kl": 0.000725990132195875,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 2.180876214057207,
-      "reward_std": 2.483732156455517,
-      "rewards/concensus_correctness_reward_func": 0.7027499993564561,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 2.022418200969696,
+      "reward_std": 2.153670124709606,
+      "rewards/concensus_correctness_reward_func": 0.7094374895095825,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.5586887625977397,
-      "rewards/soft_format_reward_func": 0.015625,
-      "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": 0.3100624904036522,
+      "rewards/question_recreation_reward_func": 0.5288244183175266,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.2685312498360872,
       "step": 14
     },
     {
-      "completion_length": 558.75,
+      "completion_length": 540.3214285714286,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.3833610713481903,
-      "kl": 0.0006043689541651734,
+      "grad_norm": 3.299898862838745,
+      "kl": 0.0007763790033225502,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 2.789768118943487,
-      "reward_std": 2.6512059484209334,
-      "rewards/concensus_correctness_reward_func": 1.4553571428571428,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 0.654516909803663,
+      "reward_std": 1.215036724294935,
+      "rewards/concensus_correctness_reward_func": 0.022142857313156128,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.2857142857142857,
-      "rewards/question_recreation_reward_func": 0.4342324904033116,
+      "rewards/final_correctness_reward_func": 0.14285714285714285,
+      "rewards/question_recreation_reward_func": 0.40419550240039825,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.11446427341018404,
+      "rewards/xmlcount_reward_func": -0.1289642857653754,
       "step": 16
     },
     {
-      "completion_length": 506.92857142857144,
+      "completion_length": 492.92857142857144,
       "epoch": 6.0,
-      "grad_norm": 0.2585050165653229,
-      "kl": 0.0006931373708149684,
+      "grad_norm": 4.3900957107543945,
+      "kl": 0.000726213857498286,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.2911256679466792,
-      "reward_std": 1.0307582169771194,
-      "rewards/concensus_correctness_reward_func": 0.02828571426549128,
-      "rewards/consensus_reward_func": 0.5714285714285714,
+      "reward": 1.6323457615716117,
+      "reward_std": 1.1951770058700018,
+      "rewards/concensus_correctness_reward_func": 0.12685714023453848,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5110899422849927,
+      "rewards/question_recreation_reward_func": 0.5524172101702008,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.18032142094203404,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.6495000209127154,
       "step": 18
     },
     {
-      "completion_length": 382.34375,
+      "completion_length": 525.34375,
       "epoch": 6.7272727272727275,
-      "grad_norm": 2.9442291259765625,
-      "kl": 0.0007867564054322429,
+      "grad_norm": 0.6854709386825562,
+      "kl": 0.0005694236060662661,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 2.731376089155674,
-      "reward_std": 2.381183974444866,
-      "rewards/concensus_correctness_reward_func": 1.2731874994933605,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 1.6998868212103844,
+      "reward_std": 2.261262886226177,
+      "rewards/concensus_correctness_reward_func": 0.7890624888241291,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.25,
-      "rewards/question_recreation_reward_func": 0.44759478233754635,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.5786056481301785,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.3074687412008643,
+      "rewards/xmlcount_reward_func": -0.1834062561392784,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 6.541969924001023e-07,
-      "train_runtime": 1061.1344,
-      "train_samples_per_second": 0.302,
-      "train_steps_per_second": 0.019
+      "train_loss": 6.380430704666651e-07,
+      "train_runtime": 1093.548,
+      "train_samples_per_second": 0.293,
+      "train_steps_per_second": 0.018
     }
   ],
   "logging_steps": 2,