End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:363c30094536b554a184c76bcd696b68cd055032abb29f7a5ed5669c70521c2a
+oid sha256:0a3c60c8633545411d906ebee87572de1c8474b074e0a529754d1e71cf3469af
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.198085086201899e-07,
-    "train_runtime": 1584.7401,
+    "train_loss": 6.581893842394493e-07,
+    "train_runtime": 1446.9107,
     "train_samples": 11,
-    "train_samples_per_second": 0.202,
-    "train_steps_per_second": 0.013
+    "train_samples_per_second": 0.221,
+    "train_steps_per_second": 0.014
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.198085086201899e-07,
-    "train_runtime": 1584.7401,
+    "train_loss": 6.581893842394493e-07,
+    "train_runtime": 1446.9107,
     "train_samples": 11,
-    "train_samples_per_second": 0.202,
-    "train_steps_per_second": 0.013
+    "train_samples_per_second": 0.221,
+    "train_steps_per_second": 0.014
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 561.59375,
+      "completion_length": 487.71875,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.3222217261791229,
-      "kl": 0.0006156760464364197,
+      "grad_norm": 0.32515591382980347,
+      "kl": 0.0005965135496808216,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 0.7097736224532127,
-      "reward_std": 1.075104609131813,
+      "reward": 1.1760332956910133,
+      "reward_std": 0.7662922292947769,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5884298821911216,
+      "rewards/question_recreation_reward_func": 0.6603145403787494,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": -0.050531258806586266,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.3907187543809414,
       "step": 2
     },
     {
-      "completion_length": 413.14285714285717,
+      "completion_length": 374.17857142857144,
       "epoch": 1.3636363636363638,
-      "grad_norm": 0.39414873719215393,
-      "kl": 0.0007785040652379394,
+      "grad_norm": 0.5938927531242371,
+      "kl": 0.0008365539833903313,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.8490615487098694,
-      "reward_std": 1.1400364807673864,
-      "rewards/concensus_correctness_reward_func": 0.0282857141324452,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 0.9771888256072998,
+      "reward_std": 0.8221025850091662,
+      "rewards/concensus_correctness_reward_func": 0.026857141937528337,
+      "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4676329472235271,
+      "rewards/question_recreation_reward_func": 0.4666173883846828,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.06742857609476362,
+      "rewards/strict_format_reward_func": 0.05357142857142857,
+      "rewards/xmlcount_reward_func": 0.28728570682661875,
       "step": 4
     },
     {
-      "completion_length": 492.4642857142857,
+      "completion_length": 409.57142857142856,
       "epoch": 2.0,
-      "grad_norm": 0.3449578285217285,
-      "kl": 0.0006918981796064015,
+      "grad_norm": 0.367230623960495,
+      "kl": 0.0007195972139015794,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 0.9568400574581963,
-      "reward_std": 0.7401026104177747,
+      "reward": 1.0640648986612047,
+      "reward_std": 0.6602635958365032,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5246257483959198,
+      "rewards/question_recreation_reward_func": 0.5395649203232357,
       "rewards/soft_format_reward_func": 0.017857142857142856,
       "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.3072142941611154,
+      "rewards/xmlcount_reward_func": 0.39950000920466017,
       "step": 6
     },
     {
-      "completion_length": 470.9375,
+      "completion_length": 431.3125,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.39589089155197144,
-      "kl": 0.0007375327404588461,
+      "grad_norm": 0.3355807662010193,
+      "kl": 0.0007665254815947264,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 0.8644137904047966,
-      "reward_std": 1.089905098080635,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.898308526724577,
+      "reward_std": 1.036042720079422,
+      "rewards/concensus_correctness_reward_func": 0.02787500061094761,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5000075455754995,
+      "rewards/question_recreation_reward_func": 0.5101522896438837,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.03125,
-      "rewards/xmlcount_reward_func": 0.02065625786781311,
+      "rewards/xmlcount_reward_func": 0.2665312569588423,
       "step": 8
     },
     {
-      "completion_length": 561.7857142857143,
+      "completion_length": 442.07142857142856,
       "epoch": 3.3636363636363638,
-      "grad_norm": 2.330106735229492,
-      "kl": 0.0006754749130258071,
+      "grad_norm": 0.4616389572620392,
+      "kl": 0.0007483451627194881,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 0.37959482840129305,
-      "reward_std": 0.7886925105537687,
+      "reward": 0.8711691583905902,
+      "reward_std": 0.8248021900653839,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.40823769356523243,
+      "rewards/question_recreation_reward_func": 0.4970620114888464,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.10007142807756152,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.2848214294229235,
       "step": 10
     },
     {
-      "completion_length": 534.4642857142857,
+      "completion_length": 470.39285714285717,
       "epoch": 4.0,
-      "grad_norm": 0.2411220520734787,
-      "kl": 0.0007088499343288797,
+      "grad_norm": 0.2698743939399719,
+      "kl": 0.0006399494720556374,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 0.9174802005290985,
-      "reward_std": 0.8755575588771275,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 1.442078743662153,
+      "reward_std": 1.0629363315446037,
+      "rewards/concensus_correctness_reward_func": 0.06371428711073739,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5952659291880471,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.07221428198473794,
+      "rewards/question_recreation_reward_func": 0.617185903447015,
+      "rewards/soft_format_reward_func": 0.017857142857142856,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.24332142301968165,
       "step": 12
     },
     {
-      "completion_length": 612.6875,
+      "completion_length": 443.40625,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.26711583137512207,
-      "kl": 0.0006393189869413618,
+      "grad_norm": 0.3577459454536438,
+      "kl": 0.0007303305101231672,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 0.5164269562810659,
-      "reward_std": 1.0301097724586725,
-      "rewards/concensus_correctness_reward_func": 0.02474999986588955,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 1.408867184072733,
+      "reward_std": 0.8001506198197603,
+      "rewards/concensus_correctness_reward_func": 0.05575000122189522,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.4383957292884588,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.498773418366909,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.1967187589034438,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.5262187458574772,
       "step": 14
     },
     {
-      "completion_length": 530.25,
+      "completion_length": 511.64285714285717,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.2758399248123169,
-      "kl": 0.0006760963130675789,
+      "grad_norm": 0.4093262255191803,
+      "kl": 0.0006703343658175852,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.7838593018906457,
-      "reward_std": 1.0056123180048806,
+      "reward": 0.6285521047455924,
+      "reward_std": 1.0692603141069412,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.07142857142857142,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5464664612497602,
+      "rewards/question_recreation_reward_func": 0.4838021035705294,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.16596427985600062,
+      "rewards/xmlcount_reward_func": -0.06953572162560054,
       "step": 16
     },
     {
-      "completion_length": 423.67857142857144,
+      "completion_length": 505.9642857142857,
       "epoch": 6.0,
-      "grad_norm": 0.3440936803817749,
-      "kl": 0.0007996462684656892,
+      "grad_norm": 0.2511747479438782,
+      "kl": 0.0006454763559824121,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 0.6838892570563725,
-      "reward_std": 0.9769647972924369,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.21428571428571427,
+      "reward": 1.4119365215301514,
+      "reward_std": 1.1672307252883911,
+      "rewards/concensus_correctness_reward_func": 0.026857141937528337,
+      "rewards/consensus_reward_func": 0.35714285714285715,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.4169606491923332,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.6560079370226178,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": -0.054499996559960504,
+      "rewards/xmlcount_reward_func": 0.33621429332665037,
       "step": 18
     },
     {
-      "completion_length": 449.3125,
+      "completion_length": 350.875,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.476371705532074,
-      "kl": 0.0014172161172609776,
+      "grad_norm": 0.4788488745689392,
+      "kl": 0.000791991529695224,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 0.9342581760138273,
-      "reward_std": 0.7589832246303558,
+      "reward": 1.052818514406681,
+      "reward_std": 0.4928256133571267,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.25,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5328519130125642,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5212247483432293,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.0732812499627471,
+      "rewards/xmlcount_reward_func": 0.5159687604755163,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 7.198085086201899e-07,
-      "train_runtime": 1584.7401,
-      "train_samples_per_second": 0.202,
-      "train_steps_per_second": 0.013
+      "train_loss": 6.581893842394493e-07,
+      "train_runtime": 1446.9107,
+      "train_samples_per_second": 0.221,
+      "train_steps_per_second": 0.014
     }
   ],
   "logging_steps": 2,