End of training

adapter_config.json

@@ -24,13 +24,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "down_proj",
-    "up_proj",
+    "gate_proj",
     "v_proj",
     "k_proj",
-    "gate_proj",
-    "o_proj"
+    "o_proj",
+    "down_proj",
+    "up_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
   "trainable_token_indices": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f62a698d541fe416a9c2e794260fcf18f758c1001a7cc217ffdad1435105bdf6
+oid sha256:03662a5e73cf75516362e8bd645a5384a3ead287560015afa95c28472eba0bae
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.693584393564379e-07,
-    "train_runtime": 972.459,
-    "train_samples": 11,
-    "train_samples_per_second": 0.329,
-    "train_steps_per_second": 0.021
+    "train_loss": 8.275917508626663e-07,
+    "train_runtime": 1309.4025,
+    "train_samples": 46,
+    "train_samples_per_second": 0.244,
+    "train_steps_per_second": 0.015
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.693584393564379e-07,
-    "train_runtime": 972.459,
-    "train_samples": 11,
-    "train_samples_per_second": 0.329,
-    "train_steps_per_second": 0.021
+    "train_loss": 8.275917508626663e-07,
+    "train_runtime": 1309.4025,
+    "train_samples": 46,
+    "train_samples_per_second": 0.244,
+    "train_steps_per_second": 0.015
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 6.7272727272727275,
+  "epoch": 1.6956521739130435,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -9,209 +9,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 560.65625,
-      "epoch": 0.7272727272727273,
-      "grad_norm": 0.42944425344467163,
-      "kl": 0.0007724234819761477,
+      "completion_length": 275.84375,
+      "epoch": 0.17391304347826086,
+      "grad_norm": 0.5431532859802246,
+      "kl": 0.0,
       "learning_rate": 4.965903258506806e-07,
       "loss": 0.0,
-      "reward": 0.8781435489654541,
-      "reward_std": 1.3415064364671707,
-      "rewards/concensus_correctness_reward_func": 0.0044999998062849045,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 2.6650467291474342,
+      "reward_std": 2.158155243843794,
+      "rewards/concensus_correctness_reward_func": 1.25,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.3940498325973749,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.5918593145906925,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.02646874636411667,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": 0.5731875039637089,
       "step": 2
     },
     {
-      "completion_length": 512.25,
-      "epoch": 1.3636363636363638,
-      "grad_norm": 0.5215281844139099,
-      "kl": 0.0007682276856420296,
+      "completion_length": 287.9375,
+      "epoch": 0.34782608695652173,
+      "grad_norm": 0.6220526099205017,
+      "kl": 0.0008928155148169026,
       "learning_rate": 4.698684378016222e-07,
       "loss": 0.0,
-      "reward": 1.3765270709991455,
-      "reward_std": 1.219215759209224,
-      "rewards/concensus_correctness_reward_func": 0.005142856921468463,
-      "rewards/consensus_reward_func": 0.35714285714285715,
+      "reward": 1.241370052099228,
+      "reward_std": 0.9243025183677673,
+      "rewards/concensus_correctness_reward_func": 0.06012500077486038,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.658705609185355,
-      "rewards/soft_format_reward_func": 0.017857142857142856,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.3198214225204928,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.46774506382644176,
+      "rewards/soft_format_reward_func": 0.015625,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.4166249968111515,
       "step": 4
     },
     {
-      "completion_length": 495.17857142857144,
-      "epoch": 2.0,
-      "grad_norm": 5.828617095947266,
-      "kl": 0.0019325501801046943,
+      "completion_length": 249.15625,
+      "epoch": 0.5217391304347826,
+      "grad_norm": 0.6066071391105652,
+      "kl": 0.0009496234270045534,
       "learning_rate": 4.193203929064353e-07,
       "loss": 0.0,
-      "reward": 1.5353988707065582,
-      "reward_std": 2.5823138696806773,
-      "rewards/concensus_correctness_reward_func": 0.7142857142857143,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 2.198715277016163,
+      "reward_std": 2.2292375192046165,
+      "rewards/concensus_correctness_reward_func": 0.9726249910891056,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.48914888873696327,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.4512778203934431,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.09660714438983373,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.32168750651180744,
       "step": 6
     },
     {
-      "completion_length": 377.53125,
-      "epoch": 2.7272727272727275,
-      "grad_norm": 0.4796436131000519,
-      "kl": 0.0008228681108448654,
+      "completion_length": 295.125,
+      "epoch": 0.6956521739130435,
+      "grad_norm": 0.5065525770187378,
+      "kl": 0.000867182548972778,
       "learning_rate": 3.5042385616324236e-07,
       "loss": 0.0,
-      "reward": 2.0229879450052977,
-      "reward_std": 2.243979901075363,
-      "rewards/concensus_correctness_reward_func": 0.6294999998062849,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 1.4596414975821972,
+      "reward_std": 2.3959164917469025,
+      "rewards/concensus_correctness_reward_func": 0.625,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.6054567517712712,
+      "rewards/final_correctness_reward_func": 0.3125,
+      "rewards/question_recreation_reward_func": 0.5142352376133204,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.28803125116974115,
+      "rewards/xmlcount_reward_func": 0.007906262762844563,
       "step": 8
     },
     {
-      "completion_length": 602.9642857142857,
-      "epoch": 3.3636363636363638,
-      "grad_norm": 0.3765483796596527,
-      "kl": 0.0006359367980621755,
+      "completion_length": 251.125,
+      "epoch": 0.8695652173913043,
+      "grad_norm": 0.9612711071968079,
+      "kl": 0.000932501679926645,
       "learning_rate": 2.706448363680831e-07,
       "loss": 0.0,
-      "reward": 0.6614550054073334,
-      "reward_std": 1.1402792717729295,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 1.1927235964685678,
+      "reward_std": 0.852789081633091,
+      "rewards/concensus_correctness_reward_func": 0.06012500077486038,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5700978658029011,
+      "rewards/final_correctness_reward_func": 0.25,
+      "rewards/question_recreation_reward_func": 0.47859859373420477,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.05357142857142857,
-      "rewards/xmlcount_reward_func": -0.1765000011239733,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.3414999907836318,
       "step": 10
     },
     {
-      "completion_length": 503.10714285714283,
-      "epoch": 4.0,
-      "grad_norm": 0.38743340969085693,
-      "kl": 0.0007121743401512504,
+      "completion_length": 272.875,
+      "epoch": 1.0,
+      "grad_norm": 0.45694512128829956,
+      "kl": 0.001253045581203575,
       "learning_rate": 1.886286282148002e-07,
       "loss": 0.0,
-      "reward": 1.1857503737722124,
-      "reward_std": 1.0749500223568507,
-      "rewards/concensus_correctness_reward_func": 0.005142856921468463,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 0.9724048500259718,
+      "reward_std": 0.8799567719300588,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.21428571428571427,
-      "rewards/question_recreation_reward_func": 0.5151789316109249,
-      "rewards/soft_format_reward_func": 0.017857142857142856,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.12971429313932145,
+      "rewards/final_correctness_reward_func": 0.3333333333333333,
+      "rewards/question_recreation_reward_func": 0.3559048945705096,
+      "rewards/soft_format_reward_func": 0.020833333333333332,
+      "rewards/strict_format_reward_func": 0.020833333333333332,
+      "rewards/xmlcount_reward_func": 0.24150000357379517,
       "step": 12
     },
     {
-      "completion_length": 470.1875,
-      "epoch": 4.7272727272727275,
-      "grad_norm": 0.3253042995929718,
-      "kl": 0.0007016189010755625,
+      "completion_length": 260.5,
+      "epoch": 1.1739130434782608,
+      "grad_norm": 0.6915552616119385,
+      "kl": 0.0010410794493509457,
       "learning_rate": 1.1326296046939333e-07,
       "loss": 0.0,
-      "reward": 0.7057941406965256,
-      "reward_std": 1.0397598631680012,
-      "rewards/concensus_correctness_reward_func": 0.008999999612569809,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 1.8954101353883743,
+      "reward_std": 2.0276321694254875,
+      "rewards/concensus_correctness_reward_func": 0.625,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4914816494565457,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.5112538626417518,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0625,
-      "rewards/xmlcount_reward_func": -0.044687503133900464,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.5404062494635582,
       "step": 14
     },
     {
-      "completion_length": 458.60714285714283,
-      "epoch": 5.363636363636363,
-      "grad_norm": 0.38039660453796387,
-      "kl": 0.0008179914563827749,
+      "completion_length": 256.875,
+      "epoch": 1.3478260869565217,
+      "grad_norm": 0.529171347618103,
+      "kl": 0.0008181983212125488,
       "learning_rate": 5.271487265090163e-08,
       "loss": 0.0,
-      "reward": 1.1329527923039027,
-      "reward_std": 1.6511083500725883,
-      "rewards/concensus_correctness_reward_func": 0.005142856921468463,
-      "rewards/consensus_reward_func": 0.42857142857142855,
+      "reward": 1.2549022659659386,
+      "reward_std": 0.8138938173651695,
+      "rewards/concensus_correctness_reward_func": 0.03125,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.6002385062830788,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.5938710477203131,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": -0.06171428518635886,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.39540625968948007,
       "step": 16
     },
     {
-      "completion_length": 476.4642857142857,
-      "epoch": 6.0,
-      "grad_norm": 0.5077833533287048,
-      "kl": 0.0006133418397179671,
+      "completion_length": 282.875,
+      "epoch": 1.5217391304347827,
+      "grad_norm": 0.6943103075027466,
+      "kl": 0.0010103160093422048,
       "learning_rate": 1.3545689574841341e-08,
       "loss": 0.0,
-      "reward": 1.3726548893111092,
-      "reward_std": 1.0581786334514618,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 2.168053776025772,
+      "reward_std": 3.060316674411297,
+      "rewards/concensus_correctness_reward_func": 1.25,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.4639763448919569,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.23010713553854398,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5551475305110216,
+      "rewards/soft_format_reward_func": 0.015625,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.3004062585532665,
       "step": 18
     },
     {
-      "completion_length": 559.5,
-      "epoch": 6.7272727272727275,
-      "grad_norm": 0.38992705941200256,
-      "kl": 0.0006704793267999776,
+      "completion_length": 316.40625,
+      "epoch": 1.6956521739130435,
+      "grad_norm": 0.47805458307266235,
+      "kl": 0.0008852209357428364,
       "learning_rate": 0.0,
       "loss": 0.0,
-      "reward": 1.2452526316046715,
-      "reward_std": 1.0222889184951782,
-      "rewards/concensus_correctness_reward_func": 0.0983749981969595,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 1.2087813690304756,
+      "reward_std": 1.2030703127384186,
+      "rewards/concensus_correctness_reward_func": 0.06012500077486038,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.5694401226937771,
+      "rewards/question_recreation_reward_func": 0.5762188471853733,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.0774375069886446,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.275562496855855,
       "step": 20
     },
     {
-      "epoch": 6.7272727272727275,
+      "epoch": 1.6956521739130435,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 7.693584393564379e-07,
-      "train_runtime": 972.459,
-      "train_samples_per_second": 0.329,
-      "train_steps_per_second": 0.021
+      "train_loss": 8.275917508626663e-07,
+      "train_runtime": 1309.4025,
+      "train_samples_per_second": 0.244,
+      "train_steps_per_second": 0.015
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 2,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {