4760

.gitattributes

@@ -51,3 +51,5 @@ global_step4420/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=l
 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
+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

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5bbdad4b95d14c3acb1e3c346a89def2fc51c362c06b818b16aae48892b9e21b
+oid sha256:19ff23fff27c77451b1866a7bb94d5615b7a1ec021c71cbe745ac754e91b428a
 size 335606536

{global_step4590 → global_step4760}/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:82c5a484d15e745c31f2c27d2a5ad405f2904a2040444a0a2d7b0923dc1cf7e6
+oid sha256:966df8bfd462ae725fadce18e8436818c1e662678fc4259a39dc5fcf82378191
 size 2013320965

{global_step4590 → global_step4760}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c6ee0675f8a2cf1a58f13aa49a13d9f87aaf3ca625123652c010c2a931779f41
+oid sha256:bd1cb96edbae032aac901439923b80dea543f4bf44a6b9ef3f416ad36cae09d0
 size 335901529

latest

@@ -1 +1 @@
-global_step4590
+global_step4760

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": 3.459016393442623,
+  "epoch": 3.5871490484266064,
   "eval_steps": 170,
-  "global_step": 4590,
+  "global_step": 4760,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -4895,6 +4895,187 @@
       "eval_samples_per_second": 5.974,
       "eval_steps_per_second": 1.495,
       "step": 4590
+    },
+    {
+      "entropy": 0.25052490234375,
+      "epoch": 3.466553608441681,
+      "grad_norm": 0.48276710510253906,
+      "learning_rate": 4.5464463979447245e-06,
+      "loss": 0.2551881313323975,
+      "mean_token_accuracy": 0.9179890528321266,
+      "num_tokens": 58399143.0,
+      "step": 4600
+    },
+    {
+      "entropy": 0.23720703125,
+      "epoch": 3.4740908234407386,
+      "grad_norm": 0.49638012051582336,
+      "learning_rate": 4.505590051615478e-06,
+      "loss": 0.24317293167114257,
+      "mean_token_accuracy": 0.9220433473587036,
+      "num_tokens": 58518939.0,
+      "step": 4610
+    },
+    {
+      "entropy": 0.23392333984375,
+      "epoch": 3.4816280384397964,
+      "grad_norm": 0.512684166431427,
+      "learning_rate": 4.4648646605139605e-06,
+      "loss": 0.23999152183532715,
+      "mean_token_accuracy": 0.9236366301774979,
+      "num_tokens": 58645386.0,
+      "step": 4620
+    },
+    {
+      "entropy": 0.2496337890625,
+      "epoch": 3.489165253438854,
+      "grad_norm": 0.476936399936676,
+      "learning_rate": 4.424271195299981e-06,
+      "loss": 0.25567328929901123,
+      "mean_token_accuracy": 0.9208142340183259,
+      "num_tokens": 58774633.0,
+      "step": 4630
+    },
+    {
+      "entropy": 0.2439453125,
+      "epoch": 3.4967024684379124,
+      "grad_norm": 0.4863242208957672,
+      "learning_rate": 4.383810623488973e-06,
+      "loss": 0.2511773109436035,
+      "mean_token_accuracy": 0.9208297878503799,
+      "num_tokens": 58890700.0,
+      "step": 4640
+    },
+    {
+      "entropy": 0.2301513671875,
+      "epoch": 3.5042396834369702,
+      "grad_norm": 0.4208335280418396,
+      "learning_rate": 4.343483909428978e-06,
+      "loss": 0.23434915542602539,
+      "mean_token_accuracy": 0.926320880651474,
+      "num_tokens": 59023930.0,
+      "step": 4650
+    },
+    {
+      "entropy": 0.2602294921875,
+      "epoch": 3.511776898436028,
+      "grad_norm": 0.5129518508911133,
+      "learning_rate": 4.3032920142776125e-06,
+      "loss": 0.2699117660522461,
+      "mean_token_accuracy": 0.9175953105092048,
+      "num_tokens": 59148853.0,
+      "step": 4660
+    },
+    {
+      "entropy": 0.232763671875,
+      "epoch": 3.519314113435086,
+      "grad_norm": 0.48988714814186096,
+      "learning_rate": 4.2632358959791984e-06,
+      "loss": 0.24585673809051514,
+      "mean_token_accuracy": 0.9236719325184822,
+      "num_tokens": 59270325.0,
+      "step": 4670
+    },
+    {
+      "entropy": 0.2343994140625,
+      "epoch": 3.5268513284341436,
+      "grad_norm": 0.5187233686447144,
+      "learning_rate": 4.2233165092419045e-06,
+      "loss": 0.23766450881958007,
+      "mean_token_accuracy": 0.9215475022792816,
+      "num_tokens": 59390235.0,
+      "step": 4680
+    },
+    {
+      "entropy": 0.2481201171875,
+      "epoch": 3.5343885434332014,
+      "grad_norm": 0.5581011772155762,
+      "learning_rate": 4.183534805515017e-06,
+      "loss": 0.2518896579742432,
+      "mean_token_accuracy": 0.9208254173398018,
+      "num_tokens": 59523333.0,
+      "step": 4690
+    },
+    {
+      "entropy": 0.2261474609375,
+      "epoch": 3.541925758432259,
+      "grad_norm": 0.4673849046230316,
+      "learning_rate": 4.143891732966233e-06,
+      "loss": 0.222815203666687,
+      "mean_token_accuracy": 0.9255363285541535,
+      "num_tokens": 59644999.0,
+      "step": 4700
+    },
+    {
+      "entropy": 0.238232421875,
+      "epoch": 3.549462973431317,
+      "grad_norm": 0.4689350724220276,
+      "learning_rate": 4.1043882364590895e-06,
+      "loss": 0.24630405902862548,
+      "mean_token_accuracy": 0.9213015034794807,
+      "num_tokens": 59768160.0,
+      "step": 4710
+    },
+    {
+      "entropy": 0.246630859375,
+      "epoch": 3.557000188430375,
+      "grad_norm": 0.5590668320655823,
+      "learning_rate": 4.065025257530425e-06,
+      "loss": 0.2526021718978882,
+      "mean_token_accuracy": 0.9192856788635254,
+      "num_tokens": 59893438.0,
+      "step": 4720
+    },
+    {
+      "entropy": 0.2509765625,
+      "epoch": 3.5645374034294326,
+      "grad_norm": 0.5854102373123169,
+      "learning_rate": 4.025803734367951e-06,
+      "loss": 0.2594055414199829,
+      "mean_token_accuracy": 0.9179863676428794,
+      "num_tokens": 60011320.0,
+      "step": 4730
+    },
+    {
+      "entropy": 0.2639892578125,
+      "epoch": 3.572074618428491,
+      "grad_norm": 0.4343920648097992,
+      "learning_rate": 3.986724601787874e-06,
+      "loss": 0.27078163623809814,
+      "mean_token_accuracy": 0.9166007101535797,
+      "num_tokens": 60132251.0,
+      "step": 4740
+    },
+    {
+      "entropy": 0.2543212890625,
+      "epoch": 3.5796118334275486,
+      "grad_norm": 0.5742672681808472,
+      "learning_rate": 3.947788791212636e-06,
+      "loss": 0.2545814037322998,
+      "mean_token_accuracy": 0.9177268758416176,
+      "num_tokens": 60250677.0,
+      "step": 4750
+    },
+    {
+      "entropy": 0.2412841796875,
+      "epoch": 3.5871490484266064,
+      "grad_norm": 0.47204774618148804,
+      "learning_rate": 3.908997230648693e-06,
+      "loss": 0.2479944944381714,
+      "mean_token_accuracy": 0.9224215567111969,
+      "num_tokens": 60384182.0,
+      "step": 4760
+    },
+    {
+      "epoch": 3.5871490484266064,
+      "eval_entropy": 0.2532833972609187,
+      "eval_loss": 0.2967984080314636,
+      "eval_mean_token_accuracy": 0.9098558257143181,
+      "eval_num_tokens": 60384182.0,
+      "eval_runtime": 444.2337,
+      "eval_samples_per_second": 5.972,
+      "eval_steps_per_second": 1.495,
+      "step": 4760
     }
   ],
   "logging_steps": 10,
@@ -4914,7 +5095,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.4156001256577434e+18,
+  "total_flos": 4.576134797612548e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null