4930

.gitattributes

@@ -53,3 +53,5 @@ global_step4590/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=l
 global_step4590/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step4760/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step4760/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step4930/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step4930/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:19ff23fff27c77451b1866a7bb94d5615b7a1ec021c71cbe745ac754e91b428a
+oid sha256:2aa67bdc0753c8b9e69a952ff4cdb2479a667cc9913e4c98dcb5c9ee131d7f81
 size 335606536

{global_step4760 → global_step4930}/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:966df8bfd462ae725fadce18e8436818c1e662678fc4259a39dc5fcf82378191
+oid sha256:e9ce9468d93830b8cd8afac27de4f6680902f25c295c7bcba0d696244dbb282a
 size 2013320965

{global_step4760 → global_step4930}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd1cb96edbae032aac901439923b80dea543f4bf44a6b9ef3f416ad36cae09d0
+oid sha256:f38a4f874881f1894cd252e47437a5e42a5444661af8726b05c2746a3600121f
 size 335901529

latest

@@ -1 +1 @@
-global_step4760
+global_step4930

trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": 3910,
   "best_metric": 0.29588836431503296,
   "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-3910",
-  "epoch": 3.5871490484266064,
+  "epoch": 3.71528170341059,
   "eval_steps": 170,
-  "global_step": 4760,
+  "global_step": 4930,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -5076,6 +5076,187 @@
       "eval_samples_per_second": 5.972,
       "eval_steps_per_second": 1.495,
       "step": 4760
+    },
+    {
+      "entropy": 0.2306640625,
+      "epoch": 3.594686263425664,
+      "grad_norm": 0.44733676314353943,
+      "learning_rate": 3.870350844664421e-06,
+      "loss": 0.23280799388885498,
+      "mean_token_accuracy": 0.9231883391737938,
+      "num_tokens": 60512836.0,
+      "step": 4770
+    },
+    {
+      "entropy": 0.22529296875,
+      "epoch": 3.602223478424722,
+      "grad_norm": 0.4727942645549774,
+      "learning_rate": 3.831850554368062e-06,
+      "loss": 0.2281050443649292,
+      "mean_token_accuracy": 0.9263469472527504,
+      "num_tokens": 60644876.0,
+      "step": 4780
+    },
+    {
+      "entropy": 0.2620361328125,
+      "epoch": 3.60976069342378,
+      "grad_norm": 0.5005683302879333,
+      "learning_rate": 3.7934972773857637e-06,
+      "loss": 0.2716717481613159,
+      "mean_token_accuracy": 0.9157996863126755,
+      "num_tokens": 60767111.0,
+      "step": 4790
+    },
+    {
+      "entropy": 0.239306640625,
+      "epoch": 3.6172979084228376,
+      "grad_norm": 0.5814849734306335,
+      "learning_rate": 3.7552919278397335e-06,
+      "loss": 0.24393236637115479,
+      "mean_token_accuracy": 0.9211369171738625,
+      "num_tokens": 60888646.0,
+      "step": 4800
+    },
+    {
+      "entropy": 0.23623046875,
+      "epoch": 3.624835123421896,
+      "grad_norm": 0.5132681727409363,
+      "learning_rate": 3.7172354163264324e-06,
+      "loss": 0.23972315788269044,
+      "mean_token_accuracy": 0.9234942555427551,
+      "num_tokens": 61006787.0,
+      "step": 4810
+    },
+    {
+      "entropy": 0.239892578125,
+      "epoch": 3.6323723384209536,
+      "grad_norm": 0.449523001909256,
+      "learning_rate": 3.6793286498948777e-06,
+      "loss": 0.250550365447998,
+      "mean_token_accuracy": 0.9213675454258918,
+      "num_tokens": 61136494.0,
+      "step": 4820
+    },
+    {
+      "entropy": 0.23447265625,
+      "epoch": 3.6399095534200114,
+      "grad_norm": 0.4843718707561493,
+      "learning_rate": 3.641572532025014e-06,
+      "loss": 0.2334545373916626,
+      "mean_token_accuracy": 0.9240896537899971,
+      "num_tokens": 61255513.0,
+      "step": 4830
+    },
+    {
+      "entropy": 0.2620361328125,
+      "epoch": 3.6474467684190692,
+      "grad_norm": 0.48772865533828735,
+      "learning_rate": 3.603967962606201e-06,
+      "loss": 0.26780409812927247,
+      "mean_token_accuracy": 0.9164189457893371,
+      "num_tokens": 61377835.0,
+      "step": 4840
+    },
+    {
+      "entropy": 0.232861328125,
+      "epoch": 3.654983983418127,
+      "grad_norm": 0.488202303647995,
+      "learning_rate": 3.566515837915738e-06,
+      "loss": 0.2365469217300415,
+      "mean_token_accuracy": 0.9234168440103531,
+      "num_tokens": 61510553.0,
+      "step": 4850
+    },
+    {
+      "entropy": 0.2384765625,
+      "epoch": 3.662521198417185,
+      "grad_norm": 0.48342105746269226,
+      "learning_rate": 3.5292170505975286e-06,
+      "loss": 0.24164836406707763,
+      "mean_token_accuracy": 0.9242537707090378,
+      "num_tokens": 61641190.0,
+      "step": 4860
+    },
+    {
+      "entropy": 0.245166015625,
+      "epoch": 3.6700584134162426,
+      "grad_norm": 0.4898669421672821,
+      "learning_rate": 3.4920724896407877e-06,
+      "loss": 0.2510417699813843,
+      "mean_token_accuracy": 0.9203757330775261,
+      "num_tokens": 61772842.0,
+      "step": 4870
+    },
+    {
+      "entropy": 0.23450927734375,
+      "epoch": 3.6775956284153004,
+      "grad_norm": 0.5533488988876343,
+      "learning_rate": 3.455083040358861e-06,
+      "loss": 0.23881711959838867,
+      "mean_token_accuracy": 0.9239112094044686,
+      "num_tokens": 61901653.0,
+      "step": 4880
+    },
+    {
+      "entropy": 0.2240234375,
+      "epoch": 3.685132843414358,
+      "grad_norm": 0.3996533155441284,
+      "learning_rate": 3.4182495843681117e-06,
+      "loss": 0.22698888778686524,
+      "mean_token_accuracy": 0.927504974603653,
+      "num_tokens": 62041188.0,
+      "step": 4890
+    },
+    {
+      "entropy": 0.24088134765625,
+      "epoch": 3.692670058413416,
+      "grad_norm": 0.49964672327041626,
+      "learning_rate": 3.3815729995669322e-06,
+      "loss": 0.24840447902679444,
+      "mean_token_accuracy": 0.9213329881429673,
+      "num_tokens": 62173168.0,
+      "step": 4900
+    },
+    {
+      "entropy": 0.22628173828125,
+      "epoch": 3.7002072734124742,
+      "grad_norm": 0.4857988953590393,
+      "learning_rate": 3.345054160114789e-06,
+      "loss": 0.22398624420166016,
+      "mean_token_accuracy": 0.9263280004262924,
+      "num_tokens": 62297740.0,
+      "step": 4910
+    },
+    {
+      "entropy": 0.246728515625,
+      "epoch": 3.707744488411532,
+      "grad_norm": 0.4899304509162903,
+      "learning_rate": 3.308693936411421e-06,
+      "loss": 0.25445635318756105,
+      "mean_token_accuracy": 0.920192101597786,
+      "num_tokens": 62418710.0,
+      "step": 4920
+    },
+    {
+      "entropy": 0.24462890625,
+      "epoch": 3.71528170341059,
+      "grad_norm": 0.4868662655353546,
+      "learning_rate": 3.2724931950760576e-06,
+      "loss": 0.24715008735656738,
+      "mean_token_accuracy": 0.9194081127643585,
+      "num_tokens": 62538535.0,
+      "step": 4930
+    },
+    {
+      "epoch": 3.71528170341059,
+      "eval_entropy": 0.25196268472326805,
+      "eval_loss": 0.2968800365924835,
+      "eval_mean_token_accuracy": 0.9099598022469555,
+      "eval_num_tokens": 62538535.0,
+      "eval_runtime": 444.76,
+      "eval_samples_per_second": 5.965,
+      "eval_steps_per_second": 1.493,
+      "step": 4930
     }
   ],
   "logging_steps": 10,
@@ -5095,7 +5276,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.576134797612548e+18,
+  "total_flos": 4.740159330748203e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null