4590

.gitattributes

@@ -49,3 +49,5 @@ global_step4250/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=l
 global_step4250/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step4420/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step4420/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step4590/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step4590/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 sha256:9b48ccc0e802b5a08229cab5f0449d03ba998ebc992df09873d3ccd63e607481
+oid sha256:5bbdad4b95d14c3acb1e3c346a89def2fc51c362c06b818b16aae48892b9e21b
 size 335606536

{global_step4420 → global_step4590}/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:e2598b206cb2aaa210c078559c3c9c1913e8c099986f41eb742966cc5e1845a5
+oid sha256:82c5a484d15e745c31f2c27d2a5ad405f2904a2040444a0a2d7b0923dc1cf7e6
 size 2013320965

{global_step4420 → global_step4590}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d4880409654c72443ecc0d77ed8e9a92ca0210f321baadb144028945964d9491
+oid sha256:c6ee0675f8a2cf1a58f13aa49a13d9f87aaf3ca625123652c010c2a931779f41
 size 335901529

latest

@@ -1 +1 @@
-global_step4420
+global_step4590

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": 3.3308837384586396,
+  "epoch": 3.459016393442623,
   "eval_steps": 170,
-  "global_step": 4420,
+  "global_step": 4590,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -4714,6 +4714,187 @@
       "eval_samples_per_second": 5.971,
       "eval_steps_per_second": 1.494,
       "step": 4420
+    },
+    {
+      "entropy": 0.24404296875,
+      "epoch": 3.3384209534576974,
+      "grad_norm": 0.5156750082969666,
+      "learning_rate": 5.260085106375577e-06,
+      "loss": 0.24577805995941163,
+      "mean_token_accuracy": 0.9231465876102447,
+      "num_tokens": 56216768.0,
+      "step": 4430
+    },
+    {
+      "entropy": 0.23583984375,
+      "epoch": 3.345958168456755,
+      "grad_norm": 0.48873215913772583,
+      "learning_rate": 5.217153965785315e-06,
+      "loss": 0.24040358066558837,
+      "mean_token_accuracy": 0.9222135201096535,
+      "num_tokens": 56331628.0,
+      "step": 4440
+    },
+    {
+      "entropy": 0.236279296875,
+      "epoch": 3.353495383455813,
+      "grad_norm": 0.6165609955787659,
+      "learning_rate": 5.174336820819241e-06,
+      "loss": 0.24129061698913573,
+      "mean_token_accuracy": 0.922450278699398,
+      "num_tokens": 56466449.0,
+      "step": 4450
+    },
+    {
+      "entropy": 0.25859375,
+      "epoch": 3.361032598454871,
+      "grad_norm": 0.42945030331611633,
+      "learning_rate": 5.131634691992576e-06,
+      "loss": 0.2604536771774292,
+      "mean_token_accuracy": 0.9179983928799629,
+      "num_tokens": 56590258.0,
+      "step": 4460
+    },
+    {
+      "entropy": 0.2478515625,
+      "epoch": 3.3685698134539286,
+      "grad_norm": 0.3764921724796295,
+      "learning_rate": 5.089048597079218e-06,
+      "loss": 0.25089526176452637,
+      "mean_token_accuracy": 0.918754868209362,
+      "num_tokens": 56712466.0,
+      "step": 4470
+    },
+    {
+      "entropy": 0.251513671875,
+      "epoch": 3.376107028452987,
+      "grad_norm": 0.5503227710723877,
+      "learning_rate": 5.046579551087469e-06,
+      "loss": 0.25963895320892333,
+      "mean_token_accuracy": 0.9187079787254333,
+      "num_tokens": 56838324.0,
+      "step": 4480
+    },
+    {
+      "entropy": 0.24443359375,
+      "epoch": 3.3836442434520446,
+      "grad_norm": 0.4642918109893799,
+      "learning_rate": 5.0042285662358715e-06,
+      "loss": 0.25160040855407717,
+      "mean_token_accuracy": 0.9221544966101647,
+      "num_tokens": 56971672.0,
+      "step": 4490
+    },
+    {
+      "entropy": 0.22529296875,
+      "epoch": 3.3911814584511024,
+      "grad_norm": 0.47102871537208557,
+      "learning_rate": 4.961996651929053e-06,
+      "loss": 0.23354432582855225,
+      "mean_token_accuracy": 0.9269132241606712,
+      "num_tokens": 57101256.0,
+      "step": 4500
+    },
+    {
+      "entropy": 0.2195068359375,
+      "epoch": 3.39871867345016,
+      "grad_norm": 0.4792303144931793,
+      "learning_rate": 4.9198848147336935e-06,
+      "loss": 0.22426846027374267,
+      "mean_token_accuracy": 0.9267282068729401,
+      "num_tokens": 57220336.0,
+      "step": 4510
+    },
+    {
+      "entropy": 0.23193359375,
+      "epoch": 3.406255888449218,
+      "grad_norm": 0.46890905499458313,
+      "learning_rate": 4.877894058354518e-06,
+      "loss": 0.23823246955871583,
+      "mean_token_accuracy": 0.9238113984465599,
+      "num_tokens": 57351375.0,
+      "step": 4520
+    },
+    {
+      "entropy": 0.259375,
+      "epoch": 3.413793103448276,
+      "grad_norm": 0.4295634925365448,
+      "learning_rate": 4.836025383610382e-06,
+      "loss": 0.26934764385223386,
+      "mean_token_accuracy": 0.9180942147970199,
+      "num_tokens": 57492948.0,
+      "step": 4530
+    },
+    {
+      "entropy": 0.23662109375,
+      "epoch": 3.4213303184473336,
+      "grad_norm": 0.4977467656135559,
+      "learning_rate": 4.794279788410408e-06,
+      "loss": 0.23803329467773438,
+      "mean_token_accuracy": 0.9239297583699226,
+      "num_tokens": 57617213.0,
+      "step": 4540
+    },
+    {
+      "entropy": 0.2344482421875,
+      "epoch": 3.4288675334463914,
+      "grad_norm": 0.5020400881767273,
+      "learning_rate": 4.752658267730218e-06,
+      "loss": 0.24145655632019042,
+      "mean_token_accuracy": 0.9231008142232895,
+      "num_tokens": 57755087.0,
+      "step": 4550
+    },
+    {
+      "entropy": 0.23726806640625,
+      "epoch": 3.436404748445449,
+      "grad_norm": 0.5047474503517151,
+      "learning_rate": 4.711161813588198e-06,
+      "loss": 0.24432265758514404,
+      "mean_token_accuracy": 0.921500913798809,
+      "num_tokens": 57881092.0,
+      "step": 4560
+    },
+    {
+      "entropy": 0.2539306640625,
+      "epoch": 3.4439419634445074,
+      "grad_norm": 0.5615915060043335,
+      "learning_rate": 4.6697914150218726e-06,
+      "loss": 0.26180734634399416,
+      "mean_token_accuracy": 0.9188700020313263,
+      "num_tokens": 58019422.0,
+      "step": 4570
+    },
+    {
+      "entropy": 0.2443115234375,
+      "epoch": 3.451479178443565,
+      "grad_norm": 0.41278159618377686,
+      "learning_rate": 4.6285480580643214e-06,
+      "loss": 0.24113664627075196,
+      "mean_token_accuracy": 0.92239079028368,
+      "num_tokens": 58144501.0,
+      "step": 4580
+    },
+    {
+      "entropy": 0.249853515625,
+      "epoch": 3.459016393442623,
+      "grad_norm": 0.48389744758605957,
+      "learning_rate": 4.587432725720687e-06,
+      "loss": 0.2561469554901123,
+      "mean_token_accuracy": 0.9191392526030541,
+      "num_tokens": 58274208.0,
+      "step": 4590
+    },
+    {
+      "epoch": 3.459016393442623,
+      "eval_entropy": 0.2534054675734187,
+      "eval_loss": 0.2973636984825134,
+      "eval_mean_token_accuracy": 0.9099213545580944,
+      "eval_num_tokens": 58274208.0,
+      "eval_runtime": 444.1047,
+      "eval_samples_per_second": 5.974,
+      "eval_steps_per_second": 1.495,
+      "step": 4590
     }
   ],
   "logging_steps": 10,
@@ -4733,7 +4914,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.24965404884992e+18,
+  "total_flos": 4.4156001256577434e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null