5270

.gitattributes

@@ -57,3 +57,5 @@ global_step4930/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=l
 global_step4930/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step5100/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step5100/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step5270/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step5270/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:ed6a83ffe91827e7f1b31c868436f6ba59ac6809719423f7e38fce21ebf777c2
+oid sha256:0f1c884c2064595682cdd6ef9427acfd0b66664cbf229f705c47d32dc16bd408
 size 335606536

{global_step5100 → global_step5270}/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:05b4de656bb9490a8c41a776e6a0cc0c6ac08458c1685e19c735193cc6dee03f
+oid sha256:d7a54cd8d510820069f5159dc7f1cae5ed23df936680ded61f2296ad4257943a
 size 2013320965

{global_step5100 → global_step5270}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b66b08c8e6ed2589033f193e75af64262f9d02ecbb86cb5911db0631cf744315
+oid sha256:eda9620cbb8af9d89528b62bc4157a86532dbac2c332d3589fde281b3366d552
 size 335901529

latest

@@ -1 +1 @@
-global_step5100
+global_step5270

trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 5100,
-  "best_metric": 0.29583480954170227,
-  "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-5100",
-  "epoch": 3.8434143583945732,
+  "best_global_step": 5270,
+  "best_metric": 0.29535341262817383,
+  "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-5270",
+  "epoch": 3.9715470133785566,
   "eval_steps": 170,
-  "global_step": 5100,
+  "global_step": 5270,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -5438,6 +5438,187 @@
       "eval_samples_per_second": 5.97,
       "eval_steps_per_second": 1.494,
       "step": 5100
+    },
+    {
+      "entropy": 0.2464111328125,
+      "epoch": 3.850951573393631,
+      "grad_norm": 0.4669792354106903,
+      "learning_rate": 2.6491161185789525e-06,
+      "loss": 0.251127290725708,
+      "mean_token_accuracy": 0.9189326271414757,
+      "num_tokens": 64792525.0,
+      "step": 5110
+    },
+    {
+      "entropy": 0.23583984375,
+      "epoch": 3.858488788392689,
+      "grad_norm": 0.5284295678138733,
+      "learning_rate": 2.6161050231688145e-06,
+      "loss": 0.23496873378753663,
+      "mean_token_accuracy": 0.9242655590176583,
+      "num_tokens": 64927953.0,
+      "step": 5120
+    },
+    {
+      "entropy": 0.2385986328125,
+      "epoch": 3.8660260033917466,
+      "grad_norm": 0.4746510982513428,
+      "learning_rate": 2.583269917475015e-06,
+      "loss": 0.24602668285369872,
+      "mean_token_accuracy": 0.9217883810400963,
+      "num_tokens": 65060416.0,
+      "step": 5130
+    },
+    {
+      "entropy": 0.2327392578125,
+      "epoch": 3.873563218390805,
+      "grad_norm": 0.5625964999198914,
+      "learning_rate": 2.5506115840981905e-06,
+      "loss": 0.23991544246673585,
+      "mean_token_accuracy": 0.9225243136286736,
+      "num_tokens": 65180637.0,
+      "step": 5140
+    },
+    {
+      "entropy": 0.25853271484375,
+      "epoch": 3.8811004333898627,
+      "grad_norm": 0.49249470233917236,
+      "learning_rate": 2.5181308014257445e-06,
+      "loss": 0.2658606290817261,
+      "mean_token_accuracy": 0.9198394268751144,
+      "num_tokens": 65306476.0,
+      "step": 5150
+    },
+    {
+      "entropy": 0.2334228515625,
+      "epoch": 3.8886376483889205,
+      "grad_norm": 0.5490977168083191,
+      "learning_rate": 2.485828343613288e-06,
+      "loss": 0.23961324691772462,
+      "mean_token_accuracy": 0.9249249547719955,
+      "num_tokens": 65429741.0,
+      "step": 5160
+    },
+    {
+      "entropy": 0.22962646484375,
+      "epoch": 3.8961748633879782,
+      "grad_norm": 0.4870103597640991,
+      "learning_rate": 2.453704980566194e-06,
+      "loss": 0.2331066608428955,
+      "mean_token_accuracy": 0.9261724725365639,
+      "num_tokens": 65557932.0,
+      "step": 5170
+    },
+    {
+      "entropy": 0.2351806640625,
+      "epoch": 3.903712078387036,
+      "grad_norm": 0.5044042468070984,
+      "learning_rate": 2.421761477921232e-06,
+      "loss": 0.24697554111480713,
+      "mean_token_accuracy": 0.9232055351138115,
+      "num_tokens": 65685595.0,
+      "step": 5180
+    },
+    {
+      "entropy": 0.2565673828125,
+      "epoch": 3.911249293386094,
+      "grad_norm": 0.5367307662963867,
+      "learning_rate": 2.3899985970283446e-06,
+      "loss": 0.262367844581604,
+      "mean_token_accuracy": 0.916471290588379,
+      "num_tokens": 65794307.0,
+      "step": 5190
+    },
+    {
+      "entropy": 0.24423828125,
+      "epoch": 3.9187865083851516,
+      "grad_norm": 0.4560658633708954,
+      "learning_rate": 2.358417094932477e-06,
+      "loss": 0.255014443397522,
+      "mean_token_accuracy": 0.9187598779797554,
+      "num_tokens": 65921865.0,
+      "step": 5200
+    },
+    {
+      "entropy": 0.24368896484375,
+      "epoch": 3.9263237233842094,
+      "grad_norm": 0.517406165599823,
+      "learning_rate": 2.3270177243555483e-06,
+      "loss": 0.25057861804962156,
+      "mean_token_accuracy": 0.9210366785526276,
+      "num_tokens": 66042392.0,
+      "step": 5210
+    },
+    {
+      "entropy": 0.259814453125,
+      "epoch": 3.9338609383832672,
+      "grad_norm": 0.4247106909751892,
+      "learning_rate": 2.2958012336785075e-06,
+      "loss": 0.2645686626434326,
+      "mean_token_accuracy": 0.9183212980628014,
+      "num_tokens": 66174262.0,
+      "step": 5220
+    },
+    {
+      "entropy": 0.24361572265625,
+      "epoch": 3.941398153382325,
+      "grad_norm": 0.4038093388080597,
+      "learning_rate": 2.2647683669234964e-06,
+      "loss": 0.24378161430358886,
+      "mean_token_accuracy": 0.9212676301598549,
+      "num_tokens": 66308633.0,
+      "step": 5230
+    },
+    {
+      "entropy": 0.239697265625,
+      "epoch": 3.948935368381383,
+      "grad_norm": 0.5646233558654785,
+      "learning_rate": 2.2339198637361094e-06,
+      "loss": 0.24792141914367677,
+      "mean_token_accuracy": 0.9222084075212479,
+      "num_tokens": 66434122.0,
+      "step": 5240
+    },
+    {
+      "entropy": 0.25,
+      "epoch": 3.956472583380441,
+      "grad_norm": 0.5821533203125,
+      "learning_rate": 2.2032564593677773e-06,
+      "loss": 0.26183009147644043,
+      "mean_token_accuracy": 0.9185786202549935,
+      "num_tokens": 66557954.0,
+      "step": 5250
+    },
+    {
+      "entropy": 0.2566162109375,
+      "epoch": 3.964009798379499,
+      "grad_norm": 0.5635284781455994,
+      "learning_rate": 2.172778884658231e-06,
+      "loss": 0.2627396821975708,
+      "mean_token_accuracy": 0.9167688429355622,
+      "num_tokens": 66677221.0,
+      "step": 5260
+    },
+    {
+      "entropy": 0.2368408203125,
+      "epoch": 3.9715470133785566,
+      "grad_norm": 0.41133439540863037,
+      "learning_rate": 2.1424878660180935e-06,
+      "loss": 0.23894243240356444,
+      "mean_token_accuracy": 0.9241284295916558,
+      "num_tokens": 66810692.0,
+      "step": 5270
+    },
+    {
+      "epoch": 3.9715470133785566,
+      "eval_entropy": 0.25422613304781627,
+      "eval_loss": 0.29535341262817383,
+      "eval_mean_token_accuracy": 0.9102872919784971,
+      "eval_num_tokens": 66810692.0,
+      "eval_runtime": 444.4912,
+      "eval_samples_per_second": 5.969,
+      "eval_steps_per_second": 1.494,
+      "step": 5270
     }
   ],
   "logging_steps": 10,
@@ -5457,7 +5638,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.900229278651646e+18,
+  "total_flos": 5.062071673121931e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null