End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:92c3ac85e7e0aaa179c8804660e83876f7969bf7259d686f07523a554ccf5471
+oid sha256:74c5c734f62ec1e91f8904738f17ac189b67b09698144b80d0d170ba71020825
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 4.836914536099357e-06,
-    "train_runtime": 470.8561,
+    "train_loss": 4.860778517468134e-06,
+    "train_runtime": 451.1811,
     "train_samples": 14,
-    "train_samples_per_second": 0.68,
-    "train_steps_per_second": 0.042
+    "train_samples_per_second": 0.709,
+    "train_steps_per_second": 0.044
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 4.836914536099357e-06,
-    "train_runtime": 470.8561,
+    "train_loss": 4.860778517468134e-06,
+    "train_runtime": 451.1811,
     "train_samples": 14,
-    "train_samples_per_second": 0.68,
-    "train_steps_per_second": 0.042
+    "train_samples_per_second": 0.709,
+    "train_steps_per_second": 0.044
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 283.90625,
+      "completion_length": 299.6875,
       "epoch": 0.5714285714285714,
-      "grad_norm": 0.7534583806991577,
-      "kl": 0.0058076647110283375,
+      "grad_norm": 0.805568277835846,
+      "kl": 0.005334832851076499,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 4.425821460783482,
-      "reward_std": 4.878054510802031,
-      "rewards/concensus_correctness_reward_func": 2.690062504261732,
-      "rewards/consensus_reward_func": 0.5625,
+      "reward": 3.464356631040573,
+      "reward_std": 3.6830832213163376,
+      "rewards/concensus_correctness_reward_func": 1.516499996650964,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.4971652301028371,
+      "rewards/final_correctness_reward_func": 0.5,
+      "rewards/question_recreation_reward_func": 0.3635440729558468,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4885937534272671,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.5686875050887465,
       "step": 2
     },
     {
-      "completion_length": 290.5,
+      "completion_length": 258.2083333333333,
       "epoch": 1.0,
-      "grad_norm": 0.34793820977211,
-      "kl": 0.004806881460050742,
+      "grad_norm": 0.35999754071235657,
+      "kl": 0.006204546041165789,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 9.30508921543757,
-      "reward_std": 8.402241498231888,
-      "rewards/concensus_correctness_reward_func": 6.7438333332538605,
-      "rewards/consensus_reward_func": 0.75,
+      "reward": 7.120826264222463,
+      "reward_std": 5.3486722360054655,
+      "rewards/concensus_correctness_reward_func": 5.256999969482422,
+      "rewards/consensus_reward_func": 0.3333333333333333,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.9166666666666666,
-      "rewards/question_recreation_reward_func": 0.5142559334635735,
-      "rewards/soft_format_reward_func": 0.020833333333333332,
-      "rewards/strict_format_reward_func": 0.020833333333333332,
-      "rewards/xmlcount_reward_func": 0.3386666675408681,
+      "rewards/final_correctness_reward_func": 0.5833333333333334,
+      "rewards/question_recreation_reward_func": 0.5174931287765503,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.4296666930119197,
       "step": 4
     },
     {
-      "completion_length": 294.25,
+      "completion_length": 330.21875,
       "epoch": 1.5714285714285714,
-      "grad_norm": 0.5963157415390015,
-      "kl": 0.0050719648716039956,
+      "grad_norm": 0.562750518321991,
+      "kl": 0.004472696280572563,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 4.436228007078171,
-      "reward_std": 4.9575640596449375,
-      "rewards/concensus_correctness_reward_func": 2.5389374997466803,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 7.63015154004097,
+      "reward_std": 6.138738967478275,
+      "rewards/concensus_correctness_reward_func": 5.345937572419643,
+      "rewards/consensus_reward_func": 0.5625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5,
-      "rewards/question_recreation_reward_func": 0.4497592533007264,
+      "rewards/final_correctness_reward_func": 0.875,
+      "rewards/question_recreation_reward_func": 0.39846412278711796,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.44753124564886093,
+      "rewards/xmlcount_reward_func": 0.448250001296401,
       "step": 6
     },
     {
-      "completion_length": 257.875,
+      "completion_length": 275.3333333333333,
       "epoch": 2.0,
-      "grad_norm": 0.41273075342178345,
-      "kl": 0.00557532188637803,
+      "grad_norm": 0.28087395429611206,
+      "kl": 0.005513112836827834,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 6.27250216404597,
-      "reward_std": 6.132579942544301,
-      "rewards/concensus_correctness_reward_func": 3.6744167308012643,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 4.998453934987386,
+      "reward_std": 5.802868373692036,
+      "rewards/concensus_correctness_reward_func": 2.7670833269755044,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.9166666666666666,
-      "rewards/question_recreation_reward_func": 0.5500854253768921,
+      "rewards/final_correctness_reward_func": 1.0,
+      "rewards/question_recreation_reward_func": 0.5352872535586357,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.6313333412011465,
+      "rewards/xmlcount_reward_func": 0.44608333706855774,
       "step": 8
     },
     {
-      "completion_length": 277.34375,
+      "completion_length": 265.625,
       "epoch": 2.571428571428571,
-      "grad_norm": 0.6800386905670166,
-      "kl": 0.006272816943237558,
+      "grad_norm": 0.5893383026123047,
+      "kl": 0.006437872449168935,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 9.082367412745953,
-      "reward_std": 7.721119299530983,
-      "rewards/concensus_correctness_reward_func": 6.571812542155385,
-      "rewards/consensus_reward_func": 0.875,
+      "reward": 4.287012085318565,
+      "reward_std": 4.268554214388132,
+      "rewards/concensus_correctness_reward_func": 2.3828750252723694,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.75,
-      "rewards/question_recreation_reward_func": 0.4784925151616335,
+      "rewards/final_correctness_reward_func": 0.5625,
+      "rewards/question_recreation_reward_func": 0.5213559083640575,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.39143750071525574,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.6327812597155571,
       "step": 10
     },
     {
-      "completion_length": 282.625,
+      "completion_length": 277.125,
       "epoch": 3.0,
-      "grad_norm": 0.4411138594150543,
-      "kl": 0.005497823314120372,
+      "grad_norm": 0.47219136357307434,
+      "kl": 0.004700874638122817,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 3.0713600317637124,
-      "reward_std": 3.116919239362081,
-      "rewards/concensus_correctness_reward_func": 1.1271666636069615,
-      "rewards/consensus_reward_func": 0.4166666666666667,
+      "reward": 3.6496654550234475,
+      "reward_std": 3.3628005584081015,
+      "rewards/concensus_correctness_reward_func": 1.438583326836427,
+      "rewards/consensus_reward_func": 0.3333333333333333,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5833333333333334,
-      "rewards/question_recreation_reward_func": 0.5399849911530813,
+      "rewards/final_correctness_reward_func": 1.0,
+      "rewards/question_recreation_reward_func": 0.37791549662748974,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4042083329210679,
+      "rewards/xmlcount_reward_func": 0.4998333305120468,
       "step": 12
     },
     {
-      "completion_length": 297.46875,
+      "completion_length": 305.65625,
       "epoch": 3.571428571428571,
-      "grad_norm": 0.5087289214134216,
-      "kl": 0.0048880781105253845,
+      "grad_norm": 0.47856101393699646,
+      "kl": 0.0058737183280754834,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 7.4388823211193085,
-      "reward_std": 7.440806720405817,
-      "rewards/concensus_correctness_reward_func": 5.073499999940395,
-      "rewards/consensus_reward_func": 0.6875,
+      "reward": 6.798029333353043,
+      "reward_std": 7.053272303193808,
+      "rewards/concensus_correctness_reward_func": 4.577624961733818,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.625,
-      "rewards/question_recreation_reward_func": 0.5125072915107012,
+      "rewards/question_recreation_reward_func": 0.5491856671869755,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.5403749991673976,
+      "rewards/xmlcount_reward_func": 0.546218741685152,
       "step": 14
     },
     {
-      "completion_length": 250.41666666666666,
+      "completion_length": 270.5,
       "epoch": 4.0,
-      "grad_norm": 0.569800078868866,
-      "kl": 0.006591112818568945,
+      "grad_norm": 0.5703155994415283,
+      "kl": 0.006019380874931812,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 3.379980981349945,
-      "reward_std": 4.598638425270717,
-      "rewards/concensus_correctness_reward_func": 1.7185833329955738,
-      "rewards/consensus_reward_func": 0.3333333333333333,
+      "reward": 4.361511866251628,
+      "reward_std": 4.572113633155823,
+      "rewards/concensus_correctness_reward_func": 1.9208333318432171,
+      "rewards/consensus_reward_func": 0.4166666666666667,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5833333333333334,
-      "rewards/question_recreation_reward_func": 0.28348104159037274,
+      "rewards/final_correctness_reward_func": 1.0833333333333333,
+      "rewards/question_recreation_reward_func": 0.3057201032837232,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.46125000591079396,
+      "rewards/xmlcount_reward_func": 0.6349583466847738,
       "step": 16
     },
     {
-      "completion_length": 281.125,
+      "completion_length": 316.96875,
       "epoch": 4.571428571428571,
-      "grad_norm": 0.7097270488739014,
-      "kl": 0.0054986039467621595,
+      "grad_norm": 0.5486003160476685,
+      "kl": 0.005313984933309257,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 4.543876305222511,
-      "reward_std": 4.88472287543118,
-      "rewards/concensus_correctness_reward_func": 2.52449999935925,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 6.939895495772362,
+      "reward_std": 6.605157300829887,
+      "rewards/concensus_correctness_reward_func": 4.545375058893114,
+      "rewards/consensus_reward_func": 0.5625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.6875,
-      "rewards/question_recreation_reward_func": 0.49393886514008045,
-      "rewards/soft_format_reward_func": 0.0,
+      "rewards/final_correctness_reward_func": 0.75,
+      "rewards/question_recreation_reward_func": 0.6286455430090427,
+      "rewards/soft_format_reward_func": 0.015625,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.462937505915761,
+      "rewards/xmlcount_reward_func": 0.4377500005066395,
       "step": 18
     },
     {
-      "completion_length": 282.0833333333333,
+      "completion_length": 282.6666666666667,
       "epoch": 5.0,
-      "grad_norm": 0.5285484790802002,
-      "kl": 0.005150872282683849,
+      "grad_norm": 0.6668877601623535,
+      "kl": 0.005708139312143127,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 6.548266510168712,
-      "reward_std": 6.881598869959514,
-      "rewards/concensus_correctness_reward_func": 4.338916666805744,
-      "rewards/consensus_reward_func": 0.75,
+      "reward": 4.106608271598816,
+      "reward_std": 4.213843494653702,
+      "rewards/concensus_correctness_reward_func": 1.9545832723379135,
+      "rewards/consensus_reward_func": 0.4166666666666667,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5833333333333334,
-      "rewards/question_recreation_reward_func": 0.4459747299551964,
+      "rewards/final_correctness_reward_func": 0.9166666666666666,
+      "rewards/question_recreation_reward_func": 0.4055250361561775,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4300416633486748,
+      "rewards/xmlcount_reward_func": 0.413166676958402,
       "step": 20
     },
     {
       "epoch": 5.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 4.836914536099357e-06,
-      "train_runtime": 470.8561,
-      "train_samples_per_second": 0.68,
-      "train_steps_per_second": 0.042
+      "train_loss": 4.860778517468134e-06,
+      "train_runtime": 451.1811,
+      "train_samples_per_second": 0.709,
+      "train_steps_per_second": 0.044
     }
   ],
   "logging_steps": 2,