End of training

adapter_config.json

@@ -24,13 +24,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "gate_proj",
-    "q_proj",
     "k_proj",
+    "q_proj",
+    "gate_proj",
     "down_proj",
     "v_proj",
-    "o_proj",
-    "up_proj"
+    "up_proj",
+    "o_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:de5ebe9418452cad9f28e2dc1b04b3adc1ba70ae1f6fe15e275893c945bfe555
+oid sha256:26c6d59dd180a88f1b989c8707760339fd57b2c2c2656011b213154aa9d8afff
 size 35237104

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.691360682798404e-06,
-    "train_runtime": 1206.6799,
-    "train_samples": 14,
-    "train_samples_per_second": 0.265,
-    "train_steps_per_second": 0.017
+    "train_loss": 2.451066382036515e-06,
+    "train_runtime": 531.9792,
+    "train_samples": 13,
+    "train_samples_per_second": 0.602,
+    "train_steps_per_second": 0.038
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.691360682798404e-06,
-    "train_runtime": 1206.6799,
-    "train_samples": 14,
-    "train_samples_per_second": 0.265,
-    "train_steps_per_second": 0.017
+    "train_loss": 2.451066382036515e-06,
+    "train_runtime": 531.9792,
+    "train_samples": 13,
+    "train_samples_per_second": 0.602,
+    "train_steps_per_second": 0.038
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 427.65625,
-      "epoch": 0.5714285714285714,
-      "grad_norm": 2.02419376373291,
-      "kl": 0.001515701886091847,
+      "completion_length": 395.25,
+      "epoch": 0.6153846153846154,
+      "grad_norm": 1.7927322387695312,
+      "kl": 0.0012580822076415643,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 0.38438029680401087,
-      "reward_std": 0.37412839848548174,
-      "rewards/concensus_correctness_reward_func": 0.0,
+      "reward": 0.5056484336964786,
+      "reward_std": 0.8683814308606088,
+      "rewards/concensus_correctness_reward_func": 0.06012500077486038,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.15269278502091765,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.26574217714369297,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.23168749827891588,
+      "rewards/xmlcount_reward_func": 0.11728124879300594,
       "step": 2
     },
     {
-      "completion_length": 331.5833333333333,
+      "completion_length": 301.15,
       "epoch": 1.0,
-      "grad_norm": 1.5866038799285889,
-      "kl": 0.0014360116717095177,
+      "grad_norm": 28.787370681762695,
+      "kl": 0.007838598499074579,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.863794689377149,
-      "reward_std": 0.9592858869582415,
-      "rewards/concensus_correctness_reward_func": 0.04633333285649618,
-      "rewards/consensus_reward_func": 0.08333333333333333,
+      "reward": 0.4848457515239716,
+      "reward_std": 0.6104163944721221,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4166666666666667,
-      "rewards/question_recreation_reward_func": 0.3307946765174468,
+      "rewards/final_correctness_reward_func": 0.1,
+      "rewards/question_recreation_reward_func": 0.211995729804039,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.013333330551783243,
+      "rewards/xmlcount_reward_func": 0.17284999787807465,
       "step": 4
     },
     {
-      "completion_length": 419.28125,
-      "epoch": 1.5714285714285714,
-      "grad_norm": 17.006370544433594,
-      "kl": 0.003698940417962149,
+      "completion_length": 366.0625,
+      "epoch": 1.6153846153846154,
+      "grad_norm": 1.4565224647521973,
+      "kl": 0.0048014514686656184,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 0.3356218687258661,
-      "reward_std": 0.5437893751077354,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0,
+      "reward": 0.5369042251259089,
+      "reward_std": 0.7701209064107388,
+      "rewards/concensus_correctness_reward_func": 0.1015625,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.29584062471985817,
-      "rewards/soft_format_reward_func": 0.015625,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.1854354883544147,
+      "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.02415626123547554,
+      "rewards/xmlcount_reward_func": 0.12490624794736505,
       "step": 6
     },
     {
-      "completion_length": 377.4166666666667,
+      "completion_length": 488.75,
       "epoch": 2.0,
-      "grad_norm": 0.8195186853408813,
-      "kl": 0.0011791597547320027,
+      "grad_norm": 1.1058247089385986,
+      "kl": 0.0015862735570408403,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.4257086689273517,
-      "reward_std": 2.380374682446321,
-      "rewards/concensus_correctness_reward_func": 0.8333333333333334,
-      "rewards/consensus_reward_func": 0.0,
+      "reward": 0.7044907063245773,
+      "reward_std": 1.334967276453972,
+      "rewards/concensus_correctness_reward_func": 0.09620000123977661,
+      "rewards/consensus_reward_func": 0.1,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.16666666666666666,
-      "rewards/question_recreation_reward_func": 0.30037524923682213,
+      "rewards/final_correctness_reward_func": 0.6,
+      "rewards/question_recreation_reward_func": 0.2195407159626484,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.12533333400885263,
+      "rewards/xmlcount_reward_func": -0.3112500309944153,
       "step": 8
     },
     {
-      "completion_length": 309.96875,
-      "epoch": 2.571428571428571,
-      "grad_norm": 60.76083755493164,
-      "kl": 0.0015269295690814033,
+      "completion_length": 464.9375,
+      "epoch": 2.6153846153846154,
+      "grad_norm": 2.566697359085083,
+      "kl": 0.0013416979272733442,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.0559864901006222,
-      "reward_std": 1.5949077871628106,
-      "rewards/concensus_correctness_reward_func": 0.625,
+      "reward": 0.562221004627645,
+      "reward_std": 0.8246445022523403,
+      "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.25401771161705256,
+      "rewards/final_correctness_reward_func": 0.25,
+      "rewards/question_recreation_reward_func": 0.27678349521011114,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.11446874449029565,
+      "rewards/xmlcount_reward_func": 0.03543749637901783,
       "step": 10
     },
     {
-      "completion_length": 453.2083333333333,
+      "completion_length": 306.25,
       "epoch": 3.0,
-      "grad_norm": 1.7316957712173462,
-      "kl": 0.0015334735799115151,
+      "grad_norm": 0.7933401465415955,
+      "kl": 0.0031318686669692397,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": -0.11685750881830852,
-      "reward_std": 0.901394876341025,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0,
+      "reward": 1.342978870868683,
+      "reward_std": 1.444445651769638,
+      "rewards/concensus_correctness_reward_func": 0.18480000495910645,
+      "rewards/consensus_reward_func": 0.2,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.21022584040959677,
+      "rewards/final_correctness_reward_func": 0.2,
+      "rewards/question_recreation_reward_func": 0.37442886233329775,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.3270833318432172,
+      "rewards/xmlcount_reward_func": 0.3837499976158142,
       "step": 12
     },
     {
-      "completion_length": 343.53125,
-      "epoch": 3.571428571428571,
-      "grad_norm": 1.9017497301101685,
-      "kl": 0.0019163806282449514,
+      "completion_length": 367.0625,
+      "epoch": 3.6153846153846154,
+      "grad_norm": 1.901016354560852,
+      "kl": 0.0015947269203024916,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.2063354100100696,
-      "reward_std": 1.56057268823497,
-      "rewards/concensus_correctness_reward_func": 0.625,
+      "reward": 0.48512101359665394,
+      "reward_std": 0.8020865241996944,
+      "rewards/concensus_correctness_reward_func": 0.0234375,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.2662728647701442,
+      "rewards/final_correctness_reward_func": 0.125,
+      "rewards/question_recreation_reward_func": 0.2969335000962019,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.12756250239908695,
+      "rewards/xmlcount_reward_func": 0.03974999859929085,
       "step": 14
     },
     {
-      "completion_length": 283.625,
+      "completion_length": 421.6,
       "epoch": 4.0,
-      "grad_norm": 1.2819938659667969,
-      "kl": 0.001800706881719331,
+      "grad_norm": 1.3564997911453247,
+      "kl": 0.004650218156166374,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.3234523143619299,
-      "reward_std": 0.5301673008749882,
+      "reward": 0.2924229323863983,
+      "reward_std": 0.7882522225379944,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.2744523075719674,
+      "rewards/question_recreation_reward_func": 0.27797292321920397,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.0489999974767367,
+      "rewards/xmlcount_reward_func": 0.01444999873638153,
       "step": 16
     },
     {
-      "completion_length": 477.03125,
-      "epoch": 4.571428571428571,
-      "grad_norm": 1.037368655204773,
-      "kl": 0.001921377333928831,
+      "completion_length": 260.75,
+      "epoch": 4.615384615384615,
+      "grad_norm": 1.8875579833984375,
+      "kl": 0.003812392649706453,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 0.45080720260739326,
-      "reward_std": 0.8174268724396825,
+      "reward": 0.4502076015342027,
+      "reward_std": 0.5057817947817966,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.23549469793215394,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.3271451264154166,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.09031250141561031,
+      "rewards/xmlcount_reward_func": 0.060562501195818186,
       "step": 18
     },
     {
-      "completion_length": 327.2083333333333,
+      "completion_length": 399.15,
       "epoch": 5.0,
-      "grad_norm": 33.87569808959961,
-      "kl": 0.002454451944989463,
+      "grad_norm": 1.042950987815857,
+      "kl": 0.001434546068776399,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 0.2128132482369741,
-      "reward_std": 0.6324757958451906,
+      "reward": 1.0424630206078291,
+      "reward_std": 0.9998365689069033,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.08333333333333333,
-      "rewards/question_recreation_reward_func": 0.17147991806268692,
+      "rewards/final_correctness_reward_func": 0.5,
+      "rewards/question_recreation_reward_func": 0.273763046041131,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.04200000315904617,
+      "rewards/xmlcount_reward_func": 0.2687000036239624,
       "step": 20
     },
     {
       "epoch": 5.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 1.691360682798404e-06,
-      "train_runtime": 1206.6799,
-      "train_samples_per_second": 0.265,
-      "train_steps_per_second": 0.017
+      "train_loss": 2.451066382036515e-06,
+      "train_runtime": 531.9792,
+      "train_samples_per_second": 0.602,
+      "train_steps_per_second": 0.038
     }
   ],
   "logging_steps": 2,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9fa1affac09a799e4713e0c997fe5f776fff47ddcdf89e0db4f58e9335fd6288
+oid sha256:4bb29e561ca780cabdf15008448af2b776ff2af74f735fdfba8a7a91fa80440c
 size 6008