End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a18417a746377a51c58bfcacb2b620cf4d5035ee1f52f2b6908fae4131eb5304
+oid sha256:c2675c31e6c0a9d5806cd99331becb7ec3510f2063f848cd341abd53366ad3a3
 size 35237104

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0001263236757949926,
-    "train_runtime": 858.8714,
-    "train_samples": 25,
-    "train_samples_per_second": 0.373,
-    "train_steps_per_second": 0.023
+    "train_loss": 0.0001925117045175284,
+    "train_runtime": 951.9535,
+    "train_samples": 16,
+    "train_samples_per_second": 0.336,
+    "train_steps_per_second": 0.021
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0001263236757949926,
-    "train_runtime": 858.8714,
-    "train_samples": 25,
-    "train_samples_per_second": 0.373,
-    "train_steps_per_second": 0.023
+    "train_loss": 0.0001925117045175284,
+    "train_runtime": 951.9535,
+    "train_samples": 16,
+    "train_samples_per_second": 0.336,
+    "train_steps_per_second": 0.021
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 2.96,
+  "epoch": 5.0,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -9,209 +9,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 550.09375,
-      "epoch": 0.32,
-      "grad_norm": 1.399573802947998,
-      "kl": 0.11836767988279462,
+      "completion_length": 342.53125,
+      "epoch": 0.5,
+      "grad_norm": 2.697580575942993,
+      "kl": 0.2045096643269062,
       "learning_rate": 4.965903258506806e-07,
-      "loss": 0.0001,
-      "reward": 1.3533679246902466,
-      "reward_std": 1.632722556591034,
-      "rewards/concensus_correctness_reward_func": 0.14168750122189522,
+      "loss": 0.0002,
+      "reward": 1.745147205889225,
+      "reward_std": 2.1825065463781357,
+      "rewards/concensus_correctness_reward_func": 0.3048749975860119,
       "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.3125,
-      "rewards/question_recreation_reward_func": 0.5010553952306509,
+      "rewards/final_correctness_reward_func": 0.5,
+      "rewards/question_recreation_reward_func": 0.4424284379929304,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": 0.17937501892447472,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.26346874982118607,
       "step": 2
     },
     {
-      "completion_length": 447.0625,
-      "epoch": 0.64,
-      "grad_norm": 17.784597396850586,
-      "kl": 0.11690637934952974,
+      "completion_length": 428.9375,
+      "epoch": 1.0,
+      "grad_norm": 2.5032296180725098,
+      "kl": 0.14022957580164075,
       "learning_rate": 4.698684378016222e-07,
       "loss": 0.0001,
-      "reward": 1.0916734039783478,
-      "reward_std": 1.3604983873665333,
+      "reward": 1.6658271923661232,
+      "reward_std": 1.2126220390200615,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.4284546338021755,
+      "rewards/final_correctness_reward_func": 0.6875,
+      "rewards/question_recreation_reward_func": 0.6038897335529327,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": 0.25696876272559166,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.35881249234080315,
       "step": 4
     },
     {
-      "completion_length": 403.375,
-      "epoch": 0.96,
-      "grad_norm": 1.9740228652954102,
-      "kl": 0.14023417280986905,
+      "completion_length": 351.15625,
+      "epoch": 1.5,
+      "grad_norm": 3.6602566242218018,
+      "kl": 0.21561020612716675,
       "learning_rate": 4.193203929064353e-07,
-      "loss": 0.0001,
-      "reward": 2.2594313360750675,
-      "reward_std": 2.2820060439407825,
-      "rewards/concensus_correctness_reward_func": 0.625,
-      "rewards/consensus_reward_func": 0.0625,
+      "loss": 0.0002,
+      "reward": 2.0754941925406456,
+      "reward_std": 1.284906916320324,
+      "rewards/concensus_correctness_reward_func": 0.18262499943375587,
+      "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5,
-      "rewards/question_recreation_reward_func": 0.5171500165015459,
+      "rewards/final_correctness_reward_func": 0.8125,
+      "rewards/question_recreation_reward_func": 0.3937129005789757,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": 0.5235312487930059,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.5616562440991402,
       "step": 6
     },
     {
-      "completion_length": 287.25,
-      "epoch": 1.16,
-      "grad_norm": 2.0795371532440186,
-      "kl": 0.16632518619298936,
+      "completion_length": 312.125,
+      "epoch": 2.0,
+      "grad_norm": 2.4019434452056885,
+      "kl": 0.18422015570104122,
       "learning_rate": 3.5042385616324236e-07,
-      "loss": 0.0001,
-      "reward": 1.753124713897705,
-      "reward_std": 1.1450692057609557,
-      "rewards/concensus_correctness_reward_func": 0.0652999997138977,
-      "rewards/consensus_reward_func": 0.1,
+      "loss": 0.0002,
+      "reward": 1.8387284837663174,
+      "reward_std": 1.264287393540144,
+      "rewards/concensus_correctness_reward_func": 0.17999999597668648,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.2,
-      "rewards/question_recreation_reward_func": 0.5341247200965882,
+      "rewards/final_correctness_reward_func": 0.3125,
+      "rewards/question_recreation_reward_func": 0.49219725653529167,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.025,
-      "rewards/xmlcount_reward_func": 0.8286999940872193,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.6509062424302101,
       "step": 8
     },
     {
-      "completion_length": 398.75,
-      "epoch": 1.48,
-      "grad_norm": 3.3047165870666504,
-      "kl": 0.1297409050166607,
+      "completion_length": 274.75,
+      "epoch": 2.5,
+      "grad_norm": 2.310741901397705,
+      "kl": 0.19709494523704052,
       "learning_rate": 2.706448363680831e-07,
-      "loss": 0.0001,
-      "reward": 1.329055480659008,
-      "reward_std": 1.5841112732887268,
-      "rewards/concensus_correctness_reward_func": 0.11993749812245369,
+      "loss": 0.0002,
+      "reward": 2.4203609600663185,
+      "reward_std": 2.9076237976551056,
+      "rewards/concensus_correctness_reward_func": 0.744999997317791,
       "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.3125,
-      "rewards/question_recreation_reward_func": 0.3880242295563221,
+      "rewards/final_correctness_reward_func": 0.5625,
+      "rewards/question_recreation_reward_func": 0.5432046614587307,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.36796876043081284,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.41340624541044235,
       "step": 10
     },
     {
-      "completion_length": 441.78125,
-      "epoch": 1.8,
-      "grad_norm": 1.552537441253662,
-      "kl": 0.13101085042580962,
+      "completion_length": 441.09375,
+      "epoch": 3.0,
+      "grad_norm": 75.53887176513672,
+      "kl": 0.22950915526598692,
       "learning_rate": 1.886286282148002e-07,
-      "loss": 0.0001,
-      "reward": 2.46834072843194,
-      "reward_std": 2.672892175614834,
-      "rewards/concensus_correctness_reward_func": 0.8032499998807907,
-      "rewards/consensus_reward_func": 0.1875,
+      "loss": 0.0002,
+      "reward": 1.5571247674524784,
+      "reward_std": 1.3080769293010235,
+      "rewards/concensus_correctness_reward_func": 0.0078125,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5625,
-      "rewards/question_recreation_reward_func": 0.5308094453066587,
+      "rewards/final_correctness_reward_func": 0.875,
+      "rewards/question_recreation_reward_func": 0.4485935387201607,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.33740624226629734,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.21009375527501106,
       "step": 12
     },
     {
-      "completion_length": 275.4,
-      "epoch": 2.0,
-      "grad_norm": 0.8664955496788025,
-      "kl": 0.160061377286911,
+      "completion_length": 288.3125,
+      "epoch": 3.5,
+      "grad_norm": 2.2316548824310303,
+      "kl": 0.23224234208464622,
       "learning_rate": 1.1326296046939333e-07,
-      "loss": 0.0001,
-      "reward": 1.430610030889511,
-      "reward_std": 1.8270696163177491,
-      "rewards/concensus_correctness_reward_func": 0.1962000012397766,
-      "rewards/consensus_reward_func": 0.1,
+      "loss": 0.0002,
+      "reward": 1.8905891850590706,
+      "reward_std": 1.0045274421572685,
+      "rewards/concensus_correctness_reward_func": 0.05999999865889549,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4,
-      "rewards/question_recreation_reward_func": 0.41111007034778596,
+      "rewards/final_correctness_reward_func": 0.5,
+      "rewards/question_recreation_reward_func": 0.4616204611957073,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.025,
-      "rewards/xmlcount_reward_func": 0.2982999861240387,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.7908437252044678,
       "step": 14
     },
     {
-      "completion_length": 376.46875,
-      "epoch": 2.32,
-      "grad_norm": 1.9497309923171997,
-      "kl": 0.14505027513951063,
+      "completion_length": 419.90625,
+      "epoch": 4.0,
+      "grad_norm": 2.349464178085327,
+      "kl": 0.15462469775229692,
       "learning_rate": 5.271487265090163e-08,
-      "loss": 0.0001,
-      "reward": 2.640411414206028,
-      "reward_std": 2.325416361913085,
-      "rewards/concensus_correctness_reward_func": 0.8051875084638596,
-      "rewards/consensus_reward_func": 0.1875,
+      "loss": 0.0002,
+      "reward": 2.160572551190853,
+      "reward_std": 2.254900734871626,
+      "rewards/concensus_correctness_reward_func": 0.625,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.524723844602704,
+      "rewards/final_correctness_reward_func": 0.625,
+      "rewards/question_recreation_reward_func": 0.4269788209348917,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.7323750033974648,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.48359374701976776,
       "step": 16
     },
     {
-      "completion_length": 462.875,
-      "epoch": 2.64,
-      "grad_norm": 1.9649162292480469,
-      "kl": 0.1616777724120766,
+      "completion_length": 319.125,
+      "epoch": 4.5,
+      "grad_norm": 2.6583991050720215,
+      "kl": 0.1754828179255128,
       "learning_rate": 1.3545689574841341e-08,
       "loss": 0.0002,
-      "reward": 1.6572420448064804,
-      "reward_std": 1.0514758825302124,
-      "rewards/concensus_correctness_reward_func": 0.08162499964237213,
-      "rewards/consensus_reward_func": 0.125,
+      "reward": 1.8738579750061035,
+      "reward_std": 0.8438492137938738,
+      "rewards/concensus_correctness_reward_func": 0.015625,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.44111706875264645,
+      "rewards/final_correctness_reward_func": 0.6875,
+      "rewards/question_recreation_reward_func": 0.4489204566925764,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.6188750080764294,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.7218125015497208,
       "step": 18
     },
     {
-      "completion_length": 486.3125,
-      "epoch": 2.96,
-      "grad_norm": 2.819248676300049,
-      "kl": 0.11614832375198603,
+      "completion_length": 381.28125,
+      "epoch": 5.0,
+      "grad_norm": 2.1007590293884277,
+      "kl": 0.19167916290462017,
       "learning_rate": 0.0,
-      "loss": 0.0001,
-      "reward": 1.4251938643865287,
-      "reward_std": 1.3764829188585281,
-      "rewards/concensus_correctness_reward_func": 0.1185000017285347,
-      "rewards/consensus_reward_func": 0.0,
+      "loss": 0.0002,
+      "reward": 1.8653074279427528,
+      "reward_std": 1.1710173524916172,
+      "rewards/concensus_correctness_reward_func": 0.05999999865889549,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.457818828523159,
+      "rewards/final_correctness_reward_func": 0.75,
+      "rewards/question_recreation_reward_func": 0.5468699478078634,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.3645000159740448,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.41468751803040504,
       "step": 20
     },
     {
-      "epoch": 2.96,
+      "epoch": 5.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 0.0001263236757949926,
-      "train_runtime": 858.8714,
-      "train_samples_per_second": 0.373,
-      "train_steps_per_second": 0.023
+      "train_loss": 0.0001925117045175284,
+      "train_runtime": 951.9535,
+      "train_samples_per_second": 0.336,
+      "train_steps_per_second": 0.021
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 5,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {