End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:81333b90ecd54d3815d0045ec9cf3f04c784d9fd969b156e666be49f68535fc8
+oid sha256:917718df5314f9dcc64d86a654c1666dc0e0bc9667c55f8ed9f5157383594f77
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.397323603669065e-06,
-    "train_runtime": 601.4621,
+    "train_loss": 1.3379344068198407e-06,
+    "train_runtime": 591.9354,
     "train_samples": 11,
-    "train_samples_per_second": 0.532,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.541,
+    "train_steps_per_second": 0.034
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.397323603669065e-06,
-    "train_runtime": 601.4621,
+    "train_loss": 1.3379344068198407e-06,
+    "train_runtime": 591.9354,
     "train_samples": 11,
-    "train_samples_per_second": 0.532,
-    "train_steps_per_second": 0.033
+    "train_samples_per_second": 0.541,
+    "train_steps_per_second": 0.034
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 492.90625,
+      "completion_length": 550.71875,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.4325464963912964,
-      "kl": 0.001182749962026719,
+      "grad_norm": 0.5056862831115723,
+      "kl": 0.0013800587548757903,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 3.0991940908133984,
-      "reward_std": 2.3954716324806213,
-      "rewards/concensus_correctness_reward_func": 1.9037500005215406,
-      "rewards/consensus_reward_func": 0.4375,
+      "reward": 1.312247209250927,
+      "reward_std": 1.1249291487038136,
+      "rewards/concensus_correctness_reward_func": 0.026562499348074198,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.528444018214941,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6732159778475761,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.16700001154094934,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.22184374369680882,
       "step": 2
     },
     {
-      "completion_length": 392.0357142857143,
+      "completion_length": 415.64285714285717,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.9016782641410828,
-      "kl": 0.002184914699422994,
+      "grad_norm": 0.4648090600967407,
+      "kl": 0.001704983363327171,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 1.0502398014068604,
-      "reward_std": 0.8396860148225512,
-      "rewards/concensus_correctness_reward_func": 0.024285714541162764,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 1.271498522588185,
+      "reward_std": 1.0002022215298243,
+      "rewards/concensus_correctness_reward_func": 0.07014285719820432,
+      "rewards/consensus_reward_func": 0.42857142857142855,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.43891836809260504,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5835699651922498,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.265607156923839,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.189214289188385,
       "step": 4
     },
     {
-      "completion_length": 458.39285714285717,
+      "completion_length": 378.5,
       "epoch": 2.0,
-      "grad_norm": 0.5998602509498596,
-      "kl": 0.001457171971976225,
+      "grad_norm": 0.6165847182273865,
+      "kl": 0.001546859059349767,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 2.6402477707181657,
-      "reward_std": 2.46128249168396,
-      "rewards/concensus_correctness_reward_func": 1.4614285720246178,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 1.2759263174874442,
+      "reward_std": 0.8719319019998822,
+      "rewards/concensus_correctness_reward_func": 0.0247142847095217,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5276406343494143,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.5256763185773577,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4368928585733686,
+      "rewards/strict_format_reward_func": 0.05357142857142857,
+      "rewards/xmlcount_reward_func": 0.314821436469044,
       "step": 6
     },
     {
-      "completion_length": 457.84375,
+      "completion_length": 529.46875,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.3934825360774994,
-      "kl": 0.0013345440092962235,
+      "grad_norm": 0.28671154379844666,
+      "kl": 0.001089718221919611,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.839316537603736,
-      "reward_std": 2.0989142823964357,
-      "rewards/concensus_correctness_reward_func": 0.6401249999180436,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 1.3038343638181686,
+      "reward_std": 1.325776332989335,
+      "rewards/concensus_correctness_reward_func": 0.11287499964237213,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.1875,
-      "rewards/question_recreation_reward_func": 0.571847852319479,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.6289281100034714,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.06484375428408384,
+      "rewards/xmlcount_reward_func": -0.00046874722465872765,
       "step": 8
     },
     {
-      "completion_length": 471.60714285714283,
+      "completion_length": 417.35714285714283,
       "epoch": 3.3636363636363638,
-      "grad_norm": 8.672999382019043,
-      "kl": 0.0020219804802244262,
+      "grad_norm": 0.568111002445221,
+      "kl": 0.001521418145525136,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.058220991066524,
-      "reward_std": 1.2709995423044478,
-      "rewards/concensus_correctness_reward_func": 0.032857143453189304,
+      "reward": 1.1384957901069097,
+      "reward_std": 1.1431427342551095,
+      "rewards/concensus_correctness_reward_func": 0.06449999979564122,
       "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5839709990790912,
+      "rewards/question_recreation_reward_func": 0.42867439133780344,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.2806785830429622,
+      "rewards/strict_format_reward_func": 0.07142857142857142,
+      "rewards/xmlcount_reward_func": 0.43103571449007305,
       "step": 10
     },
     {
-      "completion_length": 611.8928571428571,
+      "completion_length": 510.25,
       "epoch": 4.0,
-      "grad_norm": 0.3671371638774872,
-      "kl": 0.0011808994292680705,
+      "grad_norm": 0.34620562195777893,
+      "kl": 0.0014288202510215342,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.3075558521917887,
-      "reward_std": 2.3825940234320506,
-      "rewards/concensus_correctness_reward_func": 0.7321428571428571,
-      "rewards/consensus_reward_func": 0.07142857142857142,
+      "reward": 1.0430249444075994,
+      "reward_std": 1.2671920103686196,
+      "rewards/concensus_correctness_reward_func": 0.005642857402563095,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5216630186353411,
+      "rewards/question_recreation_reward_func": 0.5668106632573264,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.08910714515617915,
+      "rewards/xmlcount_reward_func": 0.04200000422341483,
       "step": 12
     },
     {
-      "completion_length": 474.0625,
+      "completion_length": 515.09375,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.4356490671634674,
-      "kl": 0.0014639904475188814,
+      "grad_norm": 0.370266854763031,
+      "kl": 0.0012138541715103202,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.5478690937161446,
-      "reward_std": 1.9978195652365685,
-      "rewards/concensus_correctness_reward_func": 0.7058125138282776,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 1.2918198108673096,
+      "reward_std": 1.2725212015211582,
+      "rewards/concensus_correctness_reward_func": 0.004937500227242708,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5866503044962883,
+      "rewards/question_recreation_reward_func": 0.6141322962939739,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.13040624372661114,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.28212498873472214,
       "step": 14
     },
     {
-      "completion_length": 573.4642857142857,
+      "completion_length": 426.9642857142857,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.5115609169006348,
-      "kl": 0.001317751469157104,
+      "grad_norm": 0.5171462297439575,
+      "kl": 0.0014096008796643997,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 1.1034282616206579,
-      "reward_std": 1.1370696468012673,
-      "rewards/concensus_correctness_reward_func": 0.028142856700079783,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 0.907790060554232,
+      "reward_std": 0.8804280289581844,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5199996829032898,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.3231428646083389,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.5585757579122271,
+      "rewards/soft_format_reward_func": 0.017857142857142856,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.11707142421177455,
       "step": 16
     },
     {
-      "completion_length": 471.17857142857144,
+      "completion_length": 478.5,
       "epoch": 6.0,
-      "grad_norm": 0.43072038888931274,
-      "kl": 0.001921076435662274,
+      "grad_norm": 3.2752087116241455,
+      "kl": 0.001824594086169132,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.1319065349442619,
-      "reward_std": 0.9992793670722416,
-      "rewards/concensus_correctness_reward_func": 0.044642857142857144,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "reward": 1.1057345696857996,
+      "reward_std": 0.8967528258051191,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5650136768817902,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.46877026770796093,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.30796427386147635,
+      "rewards/xmlcount_reward_func": 0.2798214320625578,
       "step": 18
     },
     {
-      "completion_length": 503.375,
+      "completion_length": 383.78125,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.3651694059371948,
-      "kl": 0.001272667072043987,
+      "grad_norm": 0.4814499020576477,
+      "kl": 0.0014182818413246423,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.8708339184522629,
-      "reward_std": 2.2212559394538403,
-      "rewards/concensus_correctness_reward_func": 0.6512500001117587,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 1.2847145721316338,
+      "reward_std": 0.9296081736683846,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.4375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5966151300817728,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.47596458345651627,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.23234375100582838,
+      "rewards/xmlcount_reward_func": 0.3556249951943755,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 1.397323603669065e-06,
-      "train_runtime": 601.4621,
-      "train_samples_per_second": 0.532,
-      "train_steps_per_second": 0.033
+      "train_loss": 1.3379344068198407e-06,
+      "train_runtime": 591.9354,
+      "train_samples_per_second": 0.541,
+      "train_steps_per_second": 0.034
     }
   ],
   "logging_steps": 2,