2720

.gitattributes

@@ -27,3 +27,5 @@ global_step2380/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=l
 global_step2380/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step2550/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step2550/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step2720/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step2720/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 sha256:63950ae613a7a4a682523af1e4ad036e0341cb77208e8cc5623f62173def72df
+oid sha256:1377a547a61df3e253fae7fdc28de08fb751c6561735db32918f4c8d525b0ab2
 size 335606536

{global_step2550 → global_step2720}/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9383f5b8916d86093a6eb92b0fd279dd82cf9b26b5852cf94865b4a638bf0be9
+oid sha256:f4ace3614c23e2628738aeb127757a3c39cbbe5a76d5f7197f83af4a5c8f0bd8
 size 2013320965

{global_step2550 → global_step2720}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f5a8c6c4d19245bee69f2d8df5802b850d2da679481595159d89b33dfa1c656
+oid sha256:f9fe347faa8bd24b2b578f94d3f2b4d1e8588c03109978f844f6d3d69270d193
 size 335901529

latest

@@ -1 +1 @@
-global_step2550
+global_step2720

trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": 2550,
   "best_metric": 0.3034323453903198,
   "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-2550",
-  "epoch": 1.921801394384775,
+  "epoch": 2.0497456189937817,
   "eval_steps": 170,
-  "global_step": 2550,
+  "global_step": 2720,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -2723,6 +2723,187 @@
       "eval_samples_per_second": 5.965,
       "eval_steps_per_second": 1.493,
       "step": 2550
+    },
+    {
+      "entropy": 0.2989990234375,
+      "epoch": 1.9293386093838327,
+      "grad_norm": 0.40673521161079407,
+      "learning_rate": 1.406959629222893e-05,
+      "loss": 0.3103364944458008,
+      "mean_token_accuracy": 0.9063321113586426,
+      "num_tokens": 32428504.0,
+      "step": 2560
+    },
+    {
+      "entropy": 0.2962646484375,
+      "epoch": 1.9368758243828905,
+      "grad_norm": 0.3941977322101593,
+      "learning_rate": 1.40249531664958e-05,
+      "loss": 0.30964858531951905,
+      "mean_token_accuracy": 0.9068703219294548,
+      "num_tokens": 32556407.0,
+      "step": 2570
+    },
+    {
+      "entropy": 0.2732666015625,
+      "epoch": 1.9444130393819483,
+      "grad_norm": 0.4122161567211151,
+      "learning_rate": 1.3980214108958626e-05,
+      "loss": 0.27610068321228026,
+      "mean_token_accuracy": 0.9125834852457047,
+      "num_tokens": 32679188.0,
+      "step": 2580
+    },
+    {
+      "entropy": 0.252587890625,
+      "epoch": 1.9519502543810061,
+      "grad_norm": 0.31762272119522095,
+      "learning_rate": 1.3935380185939992e-05,
+      "loss": 0.26145143508911134,
+      "mean_token_accuracy": 0.9161947175860405,
+      "num_tokens": 32806126.0,
+      "step": 2590
+    },
+    {
+      "entropy": 0.28642578125,
+      "epoch": 1.959487469380064,
+      "grad_norm": 0.37945154309272766,
+      "learning_rate": 1.3890452466023545e-05,
+      "loss": 0.29998881816864015,
+      "mean_token_accuracy": 0.908545869588852,
+      "num_tokens": 32928860.0,
+      "step": 2600
+    },
+    {
+      "entropy": 0.2949462890625,
+      "epoch": 1.967024684379122,
+      "grad_norm": 0.39361295104026794,
+      "learning_rate": 1.3845432020028511e-05,
+      "loss": 0.30800106525421145,
+      "mean_token_accuracy": 0.9073520034551621,
+      "num_tokens": 33065049.0,
+      "step": 2610
+    },
+    {
+      "entropy": 0.3162109375,
+      "epoch": 1.9745618993781797,
+      "grad_norm": 0.411302387714386,
+      "learning_rate": 1.380031992098417e-05,
+      "loss": 0.33235788345336914,
+      "mean_token_accuracy": 0.9015798211097718,
+      "num_tokens": 33199078.0,
+      "step": 2620
+    },
+    {
+      "entropy": 0.2697021484375,
+      "epoch": 1.9820991143772377,
+      "grad_norm": 0.42348551750183105,
+      "learning_rate": 1.3755117244104282e-05,
+      "loss": 0.27779905796051024,
+      "mean_token_accuracy": 0.9132746890187263,
+      "num_tokens": 33326852.0,
+      "step": 2630
+    },
+    {
+      "entropy": 0.2989501953125,
+      "epoch": 1.9896363293762955,
+      "grad_norm": 0.3523179888725281,
+      "learning_rate": 1.370982506676147e-05,
+      "loss": 0.3090573787689209,
+      "mean_token_accuracy": 0.907941748201847,
+      "num_tokens": 33465412.0,
+      "step": 2640
+    },
+    {
+      "entropy": 0.289208984375,
+      "epoch": 1.9971735443753533,
+      "grad_norm": 0.3901941180229187,
+      "learning_rate": 1.3664444468461537e-05,
+      "loss": 0.30592546463012693,
+      "mean_token_accuracy": 0.9097454324364662,
+      "num_tokens": 33594937.0,
+      "step": 2650
+    },
+    {
+      "entropy": 0.2732121394230769,
+      "epoch": 2.0045223289994345,
+      "grad_norm": 0.42783066630363464,
+      "learning_rate": 1.3618976530817727e-05,
+      "loss": 0.27013728618621824,
+      "mean_token_accuracy": 0.915503862576607,
+      "num_tokens": 33713476.0,
+      "step": 2660
+    },
+    {
+      "entropy": 0.265283203125,
+      "epoch": 2.0120595439984927,
+      "grad_norm": 0.3797246217727661,
+      "learning_rate": 1.3573422337524953e-05,
+      "loss": 0.27342140674591064,
+      "mean_token_accuracy": 0.9139280915260315,
+      "num_tokens": 33841244.0,
+      "step": 2670
+    },
+    {
+      "entropy": 0.2574951171875,
+      "epoch": 2.0195967589975505,
+      "grad_norm": 0.4094177484512329,
+      "learning_rate": 1.3527782974333978e-05,
+      "loss": 0.26642181873321535,
+      "mean_token_accuracy": 0.917819993197918,
+      "num_tokens": 33965783.0,
+      "step": 2680
+    },
+    {
+      "entropy": 0.2673828125,
+      "epoch": 2.0271339739966083,
+      "grad_norm": 0.49123308062553406,
+      "learning_rate": 1.3482059529025517e-05,
+      "loss": 0.272041392326355,
+      "mean_token_accuracy": 0.9136370241641998,
+      "num_tokens": 34084425.0,
+      "step": 2690
+    },
+    {
+      "entropy": 0.2714599609375,
+      "epoch": 2.034671188995666,
+      "grad_norm": 0.38243553042411804,
+      "learning_rate": 1.3436253091384321e-05,
+      "loss": 0.28252534866333007,
+      "mean_token_accuracy": 0.9118999525904655,
+      "num_tokens": 34206356.0,
+      "step": 2700
+    },
+    {
+      "entropy": 0.2743896484375,
+      "epoch": 2.042208403994724,
+      "grad_norm": 0.4383077919483185,
+      "learning_rate": 1.3390364753173206e-05,
+      "loss": 0.2724027633666992,
+      "mean_token_accuracy": 0.9126274734735489,
+      "num_tokens": 34326720.0,
+      "step": 2710
+    },
+    {
+      "entropy": 0.264013671875,
+      "epoch": 2.0497456189937817,
+      "grad_norm": 0.42032304406166077,
+      "learning_rate": 1.3344395608107032e-05,
+      "loss": 0.27180933952331543,
+      "mean_token_accuracy": 0.9154068097472191,
+      "num_tokens": 34449082.0,
+      "step": 2720
+    },
+    {
+      "epoch": 2.0497456189937817,
+      "eval_entropy": 0.2703239716679217,
+      "eval_loss": 0.30382418632507324,
+      "eval_mean_token_accuracy": 0.9076029673577791,
+      "eval_num_tokens": 34449082.0,
+      "eval_runtime": 444.9112,
+      "eval_samples_per_second": 5.963,
+      "eval_steps_per_second": 1.492,
+      "step": 2720
     }
   ],
   "logging_steps": 10,
@@ -2742,7 +2923,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.4469065526956524e+18,
+  "total_flos": 2.6097961573663375e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null