510

.gitattributes

@@ -1,3 +1,5 @@
 adapter_model.safetensors filter=lfs diff=lfs merge=lfs -text
 global_step340/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step340/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step510/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step510/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:de5c6c86d24a3c9af0971be6ec066e9510ec2d0d69915afa962985fc5b1dc741
+oid sha256:18e8d4cf3685889011d90bc88666d59813d7530e2f1b497f9a795e064d2daf27
 size 335606536

{global_step340 → global_step510}/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:d8bff38d1969e967e685fd911c64947c101ac7a18202aae090e58f0f1d981c52
+oid sha256:305a2f9af074492a1b2cdb6482087de5371f8c105c7a734413e9a4914c8906dc
 size 2013320965

{global_step340 → global_step510}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8fb545a54f9b706c5c5ffec3116718cff0935ef9ab6622e16c036a215d8516da
+oid sha256:6438a5e2eaa48027080a1789831fe2a9530af8d560829ee560df012eafcee001
 size 335901529

latest

@@ -1 +1 @@
-global_step340
+global_step510

trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 340,
-  "best_metric": 0.3676407039165497,
-  "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-340",
-  "epoch": 0.2562653099679668,
+  "best_global_step": 510,
+  "best_metric": 0.3505273461341858,
+  "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-510",
+  "epoch": 0.3843979649519503,
   "eval_steps": 170,
-  "global_step": 340,
+  "global_step": 510,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -370,6 +370,187 @@
       "eval_samples_per_second": 5.957,
       "eval_steps_per_second": 1.491,
       "step": 340
+    },
+    {
+      "entropy": 0.3595703125,
+      "epoch": 0.26380252496702467,
+      "grad_norm": 0.4121377468109131,
+      "learning_rate": 1.9973554384979343e-05,
+      "loss": 0.373604416847229,
+      "mean_token_accuracy": 0.892127700150013,
+      "num_tokens": 4430546.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.3835205078125,
+      "epoch": 0.2713397399660825,
+      "grad_norm": 0.4015622138977051,
+      "learning_rate": 1.9969887361333313e-05,
+      "loss": 0.3918182373046875,
+      "mean_token_accuracy": 0.8868573501706123,
+      "num_tokens": 4554384.0,
+      "step": 360
+    },
+    {
+      "entropy": 0.3498046875,
+      "epoch": 0.27887695496514037,
+      "grad_norm": 0.43255379796028137,
+      "learning_rate": 1.996598271274081e-05,
+      "loss": 0.35996294021606445,
+      "mean_token_accuracy": 0.8938629716634751,
+      "num_tokens": 4690486.0,
+      "step": 370
+    },
+    {
+      "entropy": 0.3456298828125,
+      "epoch": 0.2864141699641982,
+      "grad_norm": 0.3963714838027954,
+      "learning_rate": 1.9961840532266263e-05,
+      "loss": 0.35376391410827634,
+      "mean_token_accuracy": 0.8946138739585876,
+      "num_tokens": 4814483.0,
+      "step": 380
+    },
+    {
+      "entropy": 0.332177734375,
+      "epoch": 0.29395138496325607,
+      "grad_norm": 0.3841065764427185,
+      "learning_rate": 1.9957460918635513e-05,
+      "loss": 0.34462361335754393,
+      "mean_token_accuracy": 0.897958156466484,
+      "num_tokens": 4940434.0,
+      "step": 390
+    },
+    {
+      "entropy": 0.361767578125,
+      "epoch": 0.3014885999623139,
+      "grad_norm": 0.4300925135612488,
+      "learning_rate": 1.9952843976233428e-05,
+      "loss": 0.36802771091461184,
+      "mean_token_accuracy": 0.8923172727227211,
+      "num_tokens": 5068544.0,
+      "step": 400
+    },
+    {
+      "entropy": 0.3556640625,
+      "epoch": 0.30902581496137177,
+      "grad_norm": 0.42255330085754395,
+      "learning_rate": 1.9947989815101444e-05,
+      "loss": 0.36620967388153075,
+      "mean_token_accuracy": 0.8925080612301827,
+      "num_tokens": 5181729.0,
+      "step": 410
+    },
+    {
+      "entropy": 0.3381591796875,
+      "epoch": 0.3165630299604296,
+      "grad_norm": 0.3419656753540039,
+      "learning_rate": 1.9942898550934928e-05,
+      "loss": 0.3503819465637207,
+      "mean_token_accuracy": 0.8954922616481781,
+      "num_tokens": 5306748.0,
+      "step": 420
+    },
+    {
+      "entropy": 0.3394287109375,
+      "epoch": 0.32410024495948747,
+      "grad_norm": 0.3902963101863861,
+      "learning_rate": 1.9937570305080422e-05,
+      "loss": 0.3475761651992798,
+      "mean_token_accuracy": 0.8975138142704964,
+      "num_tokens": 5438366.0,
+      "step": 430
+    },
+    {
+      "entropy": 0.3736083984375,
+      "epoch": 0.3316374599585453,
+      "grad_norm": 0.38104942440986633,
+      "learning_rate": 1.9932005204532756e-05,
+      "loss": 0.38853695392608645,
+      "mean_token_accuracy": 0.889219282567501,
+      "num_tokens": 5576392.0,
+      "step": 440
+    },
+    {
+      "entropy": 0.3273193359375,
+      "epoch": 0.3391746749576032,
+      "grad_norm": 0.35075074434280396,
+      "learning_rate": 1.9926203381932015e-05,
+      "loss": 0.33571979999542234,
+      "mean_token_accuracy": 0.8994953021407127,
+      "num_tokens": 5701170.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.3619873046875,
+      "epoch": 0.346711889956661,
+      "grad_norm": 0.34510132670402527,
+      "learning_rate": 1.9920164975560386e-05,
+      "loss": 0.37769622802734376,
+      "mean_token_accuracy": 0.8914814174175263,
+      "num_tokens": 5834397.0,
+      "step": 460
+    },
+    {
+      "entropy": 0.37041015625,
+      "epoch": 0.3542491049557189,
+      "grad_norm": 0.3143845796585083,
+      "learning_rate": 1.9913890129338846e-05,
+      "loss": 0.3819720268249512,
+      "mean_token_accuracy": 0.891846376657486,
+      "num_tokens": 5969370.0,
+      "step": 470
+    },
+    {
+      "entropy": 0.348388671875,
+      "epoch": 0.3617863199547767,
+      "grad_norm": 0.4210856556892395,
+      "learning_rate": 1.9907378992823755e-05,
+      "loss": 0.36168179512023924,
+      "mean_token_accuracy": 0.8931615963578224,
+      "num_tokens": 6086183.0,
+      "step": 480
+    },
+    {
+      "entropy": 0.33388671875,
+      "epoch": 0.3693235349538346,
+      "grad_norm": 0.3834077715873718,
+      "learning_rate": 1.9900631721203264e-05,
+      "loss": 0.3554360866546631,
+      "mean_token_accuracy": 0.8969168856739997,
+      "num_tokens": 6209383.0,
+      "step": 490
+    },
+    {
+      "entropy": 0.3376953125,
+      "epoch": 0.3768607499528924,
+      "grad_norm": 0.3776184320449829,
+      "learning_rate": 1.9893648475293646e-05,
+      "loss": 0.35178580284118655,
+      "mean_token_accuracy": 0.8986482962965965,
+      "num_tokens": 6339234.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.359765625,
+      "epoch": 0.3843979649519503,
+      "grad_norm": 0.3891402781009674,
+      "learning_rate": 1.988642942153544e-05,
+      "loss": 0.3677093982696533,
+      "mean_token_accuracy": 0.895267216861248,
+      "num_tokens": 6467113.0,
+      "step": 510
+    },
+    {
+      "epoch": 0.3843979649519503,
+      "eval_entropy": 0.3330872317394578,
+      "eval_loss": 0.3505273461341858,
+      "eval_mean_token_accuracy": 0.8962113052067986,
+      "eval_num_tokens": 6467113.0,
+      "eval_runtime": 444.8495,
+      "eval_samples_per_second": 5.964,
+      "eval_steps_per_second": 1.493,
+      "step": 510
     }
   ],
   "logging_steps": 10,
@@ -389,7 +570,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.277167335131054e+17,
+  "total_flos": 4.933998488007475e+17,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null