Training in progress, step 1000, checkpoint

checkpoint-1000/adapter_config.json

@@ -23,13 +23,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "up_proj",
+    "o_proj",
     "down_proj",
     "gate_proj",
+    "k_proj",
     "q_proj",
     "v_proj",
-    "k_proj",
-    "o_proj"
+    "up_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

checkpoint-1000/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9a23aef6ce88119081ef5d5038f81d88398f704758e608d87caec889b3a34c22
+oid sha256:52a06b821fa1a5925b03e32d5dd8c9298e9324853570124536ddf3689add7b7a
 size 145287696

checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:21a14b53f4310f66c34597c4485826c0cac126f9f574b9cdc1c71af095db7c62
+oid sha256:de6e20afd16f1b0611fe651277c33be31c2aa588050ba52136886edb652f8f1f
 size 290833618

checkpoint-1000/tokenizer_config.json

@@ -1865,7 +1865,7 @@
   "legacy": true,
   "model_max_length": 4096,
   "pad_token": "<|endoftext|>",
-  "padding_side": "right",
+  "padding_side": "left",
   "sp_model_kwargs": {},
   "spaces_between_special_tokens": false,
   "tokenizer_class": "LlamaTokenizer",

checkpoint-1000/trainer_state.json

@@ -10,862 +10,862 @@
   "log_history": [
     {
       "epoch": 0.010869565217391304,
-      "grad_norm": 2.0792834758758545,
+      "grad_norm": 2.120478868484497,
       "learning_rate": 4.981884057971015e-05,
-      "loss": 2.3996,
+      "loss": 2.4608,
       "step": 10
     },
     {
       "epoch": 0.021739130434782608,
-      "grad_norm": 1.335732102394104,
+      "grad_norm": 1.4271776676177979,
       "learning_rate": 4.963768115942029e-05,
-      "loss": 1.25,
+      "loss": 1.2741,
       "step": 20
     },
     {
       "epoch": 0.03260869565217391,
-      "grad_norm": 1.0561087131500244,
+      "grad_norm": 0.9902746677398682,
       "learning_rate": 4.945652173913044e-05,
-      "loss": 0.9325,
+      "loss": 0.9403,
       "step": 30
     },
     {
       "epoch": 0.043478260869565216,
-      "grad_norm": 1.0983027219772339,
+      "grad_norm": 1.0730223655700684,
       "learning_rate": 4.9275362318840584e-05,
-      "loss": 0.75,
+      "loss": 0.7517,
       "step": 40
     },
     {
       "epoch": 0.05434782608695652,
-      "grad_norm": 0.8451186418533325,
+      "grad_norm": 0.8592391610145569,
       "learning_rate": 4.909420289855073e-05,
-      "loss": 0.6671,
+      "loss": 0.6695,
       "step": 50
     },
     {
       "epoch": 0.05434782608695652,
-      "eval_loss": 0.8509634733200073,
-      "eval_runtime": 10.3656,
-      "eval_samples_per_second": 46.307,
-      "eval_steps_per_second": 2.894,
+      "eval_loss": 0.8473024964332581,
+      "eval_runtime": 10.3597,
+      "eval_samples_per_second": 46.334,
+      "eval_steps_per_second": 2.896,
       "step": 50
     },
     {
       "epoch": 0.06521739130434782,
-      "grad_norm": 0.7693590521812439,
+      "grad_norm": 0.7817952632904053,
       "learning_rate": 4.891304347826087e-05,
-      "loss": 0.6309,
+      "loss": 0.6325,
       "step": 60
     },
     {
       "epoch": 0.07608695652173914,
-      "grad_norm": 0.8172665238380432,
+      "grad_norm": 0.8163846135139465,
       "learning_rate": 4.873188405797102e-05,
-      "loss": 0.5979,
+      "loss": 0.5994,
       "step": 70
     },
     {
       "epoch": 0.08695652173913043,
-      "grad_norm": 0.7351016402244568,
+      "grad_norm": 0.7358843088150024,
       "learning_rate": 4.855072463768116e-05,
-      "loss": 0.6149,
+      "loss": 0.6158,
       "step": 80
     },
     {
       "epoch": 0.09782608695652174,
-      "grad_norm": 0.8038097023963928,
+      "grad_norm": 0.7856508493423462,
       "learning_rate": 4.836956521739131e-05,
-      "loss": 0.5969,
+      "loss": 0.5983,
       "step": 90
     },
     {
       "epoch": 0.10869565217391304,
-      "grad_norm": 0.8729331493377686,
+      "grad_norm": 0.8615956902503967,
       "learning_rate": 4.818840579710145e-05,
-      "loss": 0.5847,
+      "loss": 0.5858,
       "step": 100
     },
     {
       "epoch": 0.10869565217391304,
-      "eval_loss": 0.7441961765289307,
-      "eval_runtime": 10.4341,
-      "eval_samples_per_second": 46.003,
-      "eval_steps_per_second": 2.875,
+      "eval_loss": 0.7398766875267029,
+      "eval_runtime": 10.4317,
+      "eval_samples_per_second": 46.014,
+      "eval_steps_per_second": 2.876,
       "step": 100
     },
     {
       "epoch": 0.11956521739130435,
-      "grad_norm": 0.7882421016693115,
+      "grad_norm": 0.7915664315223694,
       "learning_rate": 4.80072463768116e-05,
-      "loss": 0.622,
+      "loss": 0.6228,
       "step": 110
     },
     {
       "epoch": 0.13043478260869565,
-      "grad_norm": 0.8385023474693298,
+      "grad_norm": 0.8432077765464783,
       "learning_rate": 4.782608695652174e-05,
-      "loss": 0.5637,
+      "loss": 0.5649,
       "step": 120
     },
     {
       "epoch": 0.14130434782608695,
-      "grad_norm": 0.8384169340133667,
+      "grad_norm": 0.8230385780334473,
       "learning_rate": 4.764492753623189e-05,
-      "loss": 0.5363,
+      "loss": 0.5368,
       "step": 130
     },
     {
       "epoch": 0.15217391304347827,
-      "grad_norm": 0.8619622588157654,
+      "grad_norm": 0.8497728109359741,
       "learning_rate": 4.746376811594203e-05,
-      "loss": 0.544,
+      "loss": 0.5447,
       "step": 140
     },
     {
       "epoch": 0.16304347826086957,
-      "grad_norm": 0.9462847709655762,
+      "grad_norm": 0.9394277930259705,
       "learning_rate": 4.7282608695652177e-05,
-      "loss": 0.5512,
+      "loss": 0.5515,
       "step": 150
     },
     {
       "epoch": 0.16304347826086957,
-      "eval_loss": 0.6991452574729919,
-      "eval_runtime": 10.4072,
-      "eval_samples_per_second": 46.122,
-      "eval_steps_per_second": 2.883,
+      "eval_loss": 0.694922149181366,
+      "eval_runtime": 10.4348,
+      "eval_samples_per_second": 46.0,
+      "eval_steps_per_second": 2.875,
       "step": 150
     },
     {
       "epoch": 0.17391304347826086,
-      "grad_norm": 0.7737827897071838,
+      "grad_norm": 0.7665875554084778,
       "learning_rate": 4.710144927536232e-05,
-      "loss": 0.5623,
+      "loss": 0.5632,
       "step": 160
     },
     {
       "epoch": 0.18478260869565216,
-      "grad_norm": 0.8910329937934875,
+      "grad_norm": 0.8912480473518372,
       "learning_rate": 4.6920289855072464e-05,
-      "loss": 0.5254,
+      "loss": 0.5259,
       "step": 170
     },
     {
       "epoch": 0.1956521739130435,
-      "grad_norm": 0.8883982300758362,
+      "grad_norm": 0.8959771990776062,
       "learning_rate": 4.673913043478261e-05,
-      "loss": 0.5302,
+      "loss": 0.5307,
       "step": 180
     },
     {
       "epoch": 0.20652173913043478,
-      "grad_norm": 0.855445921421051,
+      "grad_norm": 0.8595960140228271,
       "learning_rate": 4.655797101449276e-05,
-      "loss": 0.5157,
+      "loss": 0.5167,
       "step": 190
     },
     {
       "epoch": 0.21739130434782608,
-      "grad_norm": 0.8524109721183777,
+      "grad_norm": 0.8418941497802734,
       "learning_rate": 4.63768115942029e-05,
-      "loss": 0.5102,
+      "loss": 0.5106,
       "step": 200
     },
     {
       "epoch": 0.21739130434782608,
-      "eval_loss": 0.6746214628219604,
-      "eval_runtime": 10.3921,
-      "eval_samples_per_second": 46.189,
-      "eval_steps_per_second": 2.887,
+      "eval_loss": 0.6705857515335083,
+      "eval_runtime": 10.4259,
+      "eval_samples_per_second": 46.039,
+      "eval_steps_per_second": 2.877,
       "step": 200
     },
     {
       "epoch": 0.22826086956521738,
-      "grad_norm": 0.9038218855857849,
+      "grad_norm": 0.8937019109725952,
       "learning_rate": 4.6195652173913046e-05,
-      "loss": 0.5321,
+      "loss": 0.5329,
       "step": 210
     },
     {
       "epoch": 0.2391304347826087,
-      "grad_norm": 0.9139806628227234,
+      "grad_norm": 0.9121237397193909,
       "learning_rate": 4.601449275362319e-05,
-      "loss": 0.5337,
+      "loss": 0.5341,
       "step": 220
     },
     {
       "epoch": 0.25,
-      "grad_norm": 0.8306450247764587,
+      "grad_norm": 0.8253990411758423,
       "learning_rate": 4.5833333333333334e-05,
-      "loss": 0.5324,
+      "loss": 0.5328,
       "step": 230
     },
     {
       "epoch": 0.2608695652173913,
-      "grad_norm": 0.9253225326538086,
+      "grad_norm": 0.9218540787696838,
       "learning_rate": 4.565217391304348e-05,
-      "loss": 0.4702,
+      "loss": 0.4706,
       "step": 240
     },
     {
       "epoch": 0.2717391304347826,
-      "grad_norm": 0.9208872318267822,
+      "grad_norm": 0.9071463346481323,
       "learning_rate": 4.547101449275363e-05,
       "loss": 0.5012,
       "step": 250
     },
     {
       "epoch": 0.2717391304347826,
-      "eval_loss": 0.6523857116699219,
-      "eval_runtime": 10.3866,
-      "eval_samples_per_second": 46.214,
-      "eval_steps_per_second": 2.888,
+      "eval_loss": 0.6487293839454651,
+      "eval_runtime": 10.4286,
+      "eval_samples_per_second": 46.027,
+      "eval_steps_per_second": 2.877,
       "step": 250
     },
     {
       "epoch": 0.2826086956521739,
-      "grad_norm": 0.9187518954277039,
+      "grad_norm": 0.9131689071655273,
       "learning_rate": 4.528985507246377e-05,
-      "loss": 0.4811,
+      "loss": 0.4819,
       "step": 260
     },
     {
       "epoch": 0.29347826086956524,
-      "grad_norm": 0.849318265914917,
+      "grad_norm": 0.8460387587547302,
       "learning_rate": 4.5108695652173916e-05,
-      "loss": 0.494,
+      "loss": 0.4951,
       "step": 270
     },
     {
       "epoch": 0.30434782608695654,
-      "grad_norm": 0.8616385459899902,
+      "grad_norm": 0.8639910817146301,
       "learning_rate": 4.492753623188406e-05,
-      "loss": 0.4934,
+      "loss": 0.4941,
       "step": 280
     },
     {
       "epoch": 0.31521739130434784,
-      "grad_norm": 1.0425682067871094,
+      "grad_norm": 1.0238360166549683,
       "learning_rate": 4.4746376811594203e-05,
-      "loss": 0.4835,
+      "loss": 0.4844,
       "step": 290
     },
     {
       "epoch": 0.32608695652173914,
-      "grad_norm": 0.8278887271881104,
+      "grad_norm": 0.8217597603797913,
       "learning_rate": 4.456521739130435e-05,
-      "loss": 0.4961,
+      "loss": 0.497,
       "step": 300
     },
     {
       "epoch": 0.32608695652173914,
-      "eval_loss": 0.639061689376831,
-      "eval_runtime": 10.3771,
-      "eval_samples_per_second": 46.256,
-      "eval_steps_per_second": 2.891,
+      "eval_loss": 0.6351093053817749,
+      "eval_runtime": 10.4278,
+      "eval_samples_per_second": 46.031,
+      "eval_steps_per_second": 2.877,
       "step": 300
     },
     {
       "epoch": 0.33695652173913043,
-      "grad_norm": 0.8740423321723938,
+      "grad_norm": 0.8742851614952087,
       "learning_rate": 4.438405797101449e-05,
-      "loss": 0.5131,
+      "loss": 0.5134,
       "step": 310
     },
     {
       "epoch": 0.34782608695652173,
-      "grad_norm": 0.8640497326850891,
+      "grad_norm": 0.8606206774711609,
       "learning_rate": 4.4202898550724645e-05,
-      "loss": 0.4576,
+      "loss": 0.4577,
       "step": 320
     },
     {
       "epoch": 0.358695652173913,
-      "grad_norm": 1.0350477695465088,
+      "grad_norm": 1.030929446220398,
       "learning_rate": 4.4021739130434786e-05,
-      "loss": 0.4621,
+      "loss": 0.4624,
       "step": 330
     },
     {
       "epoch": 0.3695652173913043,
-      "grad_norm": 0.8288527727127075,
+      "grad_norm": 0.8211489915847778,
       "learning_rate": 4.384057971014493e-05,
-      "loss": 0.4929,
+      "loss": 0.4942,
       "step": 340
     },
     {
       "epoch": 0.3804347826086957,
-      "grad_norm": 1.0214691162109375,
+      "grad_norm": 1.0272365808486938,
       "learning_rate": 4.365942028985507e-05,
       "loss": 0.4923,
       "step": 350
     },
     {
       "epoch": 0.3804347826086957,
-      "eval_loss": 0.6334717869758606,
-      "eval_runtime": 10.3968,
-      "eval_samples_per_second": 46.168,
-      "eval_steps_per_second": 2.885,
+      "eval_loss": 0.6294742226600647,
+      "eval_runtime": 10.439,
+      "eval_samples_per_second": 45.981,
+      "eval_steps_per_second": 2.874,
       "step": 350
     },
     {
       "epoch": 0.391304347826087,
-      "grad_norm": 0.9189789295196533,
+      "grad_norm": 0.9275714159011841,
       "learning_rate": 4.347826086956522e-05,
-      "loss": 0.4781,
+      "loss": 0.4784,
       "step": 360
     },
     {
       "epoch": 0.40217391304347827,
-      "grad_norm": 0.9531463980674744,
+      "grad_norm": 0.9278607368469238,
       "learning_rate": 4.329710144927536e-05,
-      "loss": 0.4651,
+      "loss": 0.465,
       "step": 370
     },
     {
       "epoch": 0.41304347826086957,
-      "grad_norm": 1.0185341835021973,
+      "grad_norm": 1.009921669960022,
       "learning_rate": 4.3115942028985515e-05,
-      "loss": 0.477,
+      "loss": 0.4776,
       "step": 380
     },
     {
       "epoch": 0.42391304347826086,
-      "grad_norm": 0.9705609679222107,
+      "grad_norm": 0.9694483280181885,
       "learning_rate": 4.2934782608695655e-05,
-      "loss": 0.4954,
+      "loss": 0.4956,
       "step": 390
     },
     {
       "epoch": 0.43478260869565216,
-      "grad_norm": 0.8557504415512085,
+      "grad_norm": 0.8650519251823425,
       "learning_rate": 4.27536231884058e-05,
-      "loss": 0.4602,
+      "loss": 0.4607,
       "step": 400
     },
     {
       "epoch": 0.43478260869565216,
-      "eval_loss": 0.6238241791725159,
-      "eval_runtime": 10.4192,
-      "eval_samples_per_second": 46.069,
+      "eval_loss": 0.6202508807182312,
+      "eval_runtime": 10.4217,
+      "eval_samples_per_second": 46.058,
       "eval_steps_per_second": 2.879,
       "step": 400
     },
     {
       "epoch": 0.44565217391304346,
-      "grad_norm": 1.0175505876541138,
+      "grad_norm": 1.0009912252426147,
       "learning_rate": 4.257246376811594e-05,
-      "loss": 0.4682,
+      "loss": 0.4695,
       "step": 410
     },
     {
       "epoch": 0.45652173913043476,
-      "grad_norm": 0.9760918617248535,
+      "grad_norm": 0.9651597738265991,
       "learning_rate": 4.239130434782609e-05,
-      "loss": 0.4731,
+      "loss": 0.4733,
       "step": 420
     },
     {
       "epoch": 0.4673913043478261,
-      "grad_norm": 0.9057648181915283,
+      "grad_norm": 0.9021621346473694,
       "learning_rate": 4.221014492753623e-05,
-      "loss": 0.4537,
+      "loss": 0.4543,
       "step": 430
     },
     {
       "epoch": 0.4782608695652174,
-      "grad_norm": 0.9714989066123962,
+      "grad_norm": 0.9730608463287354,
       "learning_rate": 4.202898550724638e-05,
-      "loss": 0.446,
+      "loss": 0.4461,
       "step": 440
     },
     {
       "epoch": 0.4891304347826087,
-      "grad_norm": 0.950578510761261,
+      "grad_norm": 0.9395557641983032,
       "learning_rate": 4.1847826086956525e-05,
-      "loss": 0.4444,
+      "loss": 0.4448,
       "step": 450
     },
     {
       "epoch": 0.4891304347826087,
-      "eval_loss": 0.6164894104003906,
-      "eval_runtime": 10.3782,
-      "eval_samples_per_second": 46.251,
-      "eval_steps_per_second": 2.891,
+      "eval_loss": 0.6122747659683228,
+      "eval_runtime": 10.4358,
+      "eval_samples_per_second": 45.995,
+      "eval_steps_per_second": 2.875,
       "step": 450
     },
     {
       "epoch": 0.5,
-      "grad_norm": 0.9109216928482056,
+      "grad_norm": 0.9062746167182922,
       "learning_rate": 4.166666666666667e-05,
-      "loss": 0.465,
+      "loss": 0.4653,
       "step": 460
     },
     {
       "epoch": 0.5108695652173914,
-      "grad_norm": 0.9473273158073425,
+      "grad_norm": 0.9474561214447021,
       "learning_rate": 4.148550724637681e-05,
-      "loss": 0.458,
+      "loss": 0.4571,
       "step": 470
     },
     {
       "epoch": 0.5217391304347826,
-      "grad_norm": 0.921631932258606,
+      "grad_norm": 0.9197052717208862,
       "learning_rate": 4.130434782608696e-05,
       "loss": 0.4416,
       "step": 480
     },
     {
       "epoch": 0.532608695652174,
-      "grad_norm": 0.994552493095398,
+      "grad_norm": 0.9986901879310608,
       "learning_rate": 4.11231884057971e-05,
-      "loss": 0.4441,
+      "loss": 0.4446,
       "step": 490
     },
     {
       "epoch": 0.5434782608695652,
-      "grad_norm": 0.9409403800964355,
+      "grad_norm": 0.9315592050552368,
       "learning_rate": 4.094202898550725e-05,
-      "loss": 0.45,
+      "loss": 0.4501,
       "step": 500
     },
     {
       "epoch": 0.5434782608695652,
-      "eval_loss": 0.6097907423973083,
-      "eval_runtime": 10.3702,
-      "eval_samples_per_second": 46.286,
-      "eval_steps_per_second": 2.893,
+      "eval_loss": 0.6061260104179382,
+      "eval_runtime": 10.4306,
+      "eval_samples_per_second": 46.018,
+      "eval_steps_per_second": 2.876,
       "step": 500
     },
     {
       "epoch": 0.5543478260869565,
-      "grad_norm": 0.9720210433006287,
+      "grad_norm": 0.9725834131240845,
       "learning_rate": 4.076086956521739e-05,
-      "loss": 0.4468,
+      "loss": 0.4465,
       "step": 510
     },
     {
       "epoch": 0.5652173913043478,
-      "grad_norm": 0.9678219556808472,
+      "grad_norm": 0.9681097865104675,
       "learning_rate": 4.057971014492754e-05,
-      "loss": 0.4741,
+      "loss": 0.4735,
       "step": 520
     },
     {
       "epoch": 0.5760869565217391,
-      "grad_norm": 0.9966796636581421,
+      "grad_norm": 1.0040682554244995,
       "learning_rate": 4.039855072463768e-05,
-      "loss": 0.4742,
+      "loss": 0.4743,
       "step": 530
     },
     {
       "epoch": 0.5869565217391305,
-      "grad_norm": 0.9896467328071594,
+      "grad_norm": 0.9864315986633301,
       "learning_rate": 4.021739130434783e-05,
-      "loss": 0.4499,
+      "loss": 0.45,
       "step": 540
     },
     {
       "epoch": 0.5978260869565217,
-      "grad_norm": 0.9556202292442322,
+      "grad_norm": 0.954130232334137,
       "learning_rate": 4.003623188405797e-05,
-      "loss": 0.4398,
+      "loss": 0.439,
       "step": 550
     },
     {
       "epoch": 0.5978260869565217,
-      "eval_loss": 0.6025962829589844,
-      "eval_runtime": 10.3801,
-      "eval_samples_per_second": 46.242,
-      "eval_steps_per_second": 2.89,
+      "eval_loss": 0.5990516543388367,
+      "eval_runtime": 10.427,
+      "eval_samples_per_second": 46.035,
+      "eval_steps_per_second": 2.877,
       "step": 550
     },
     {
       "epoch": 0.6086956521739131,
-      "grad_norm": 1.0646259784698486,
+      "grad_norm": 1.150810718536377,
       "learning_rate": 3.985507246376812e-05,
-      "loss": 0.4596,
+      "loss": 0.4603,
       "step": 560
     },
     {
       "epoch": 0.6195652173913043,
-      "grad_norm": 0.9542304873466492,
+      "grad_norm": 0.9530227184295654,
       "learning_rate": 3.9673913043478264e-05,
       "loss": 0.447,
       "step": 570
     },
     {
       "epoch": 0.6304347826086957,
-      "grad_norm": 1.153639316558838,
+      "grad_norm": 1.1645703315734863,
       "learning_rate": 3.9492753623188405e-05,
-      "loss": 0.4194,
+      "loss": 0.4195,
       "step": 580
     },
     {
       "epoch": 0.6413043478260869,
-      "grad_norm": 1.0100998878479004,
+      "grad_norm": 1.0095150470733643,
       "learning_rate": 3.931159420289855e-05,
-      "loss": 0.4344,
+      "loss": 0.434,
       "step": 590
     },
     {
       "epoch": 0.6521739130434783,
-      "grad_norm": 0.8966252207756042,
+      "grad_norm": 0.8818724155426025,
       "learning_rate": 3.91304347826087e-05,
-      "loss": 0.4387,
+      "loss": 0.4385,
       "step": 600
     },
     {
       "epoch": 0.6521739130434783,
-      "eval_loss": 0.5988693237304688,
-      "eval_runtime": 10.3834,
-      "eval_samples_per_second": 46.228,
-      "eval_steps_per_second": 2.889,
+      "eval_loss": 0.5951303839683533,
+      "eval_runtime": 10.4035,
+      "eval_samples_per_second": 46.138,
+      "eval_steps_per_second": 2.884,
       "step": 600
     },
     {
       "epoch": 0.6630434782608695,
-      "grad_norm": 1.0731533765792847,
+      "grad_norm": 1.0614160299301147,
       "learning_rate": 3.8949275362318846e-05,
-      "loss": 0.4127,
+      "loss": 0.4128,
       "step": 610
     },
     {
       "epoch": 0.6739130434782609,
-      "grad_norm": 0.8809463977813721,
+      "grad_norm": 0.883694589138031,
       "learning_rate": 3.876811594202899e-05,
-      "loss": 0.4663,
+      "loss": 0.4664,
       "step": 620
     },
     {
       "epoch": 0.6847826086956522,
-      "grad_norm": 0.9806007742881775,
+      "grad_norm": 0.97767573595047,
       "learning_rate": 3.8586956521739134e-05,
-      "loss": 0.4146,
+      "loss": 0.4149,
       "step": 630
     },
     {
       "epoch": 0.6956521739130435,
-      "grad_norm": 0.9644826650619507,
+      "grad_norm": 0.9714617133140564,
       "learning_rate": 3.8405797101449274e-05,
-      "loss": 0.4378,
+      "loss": 0.4369,
       "step": 640
     },
     {
       "epoch": 0.7065217391304348,
-      "grad_norm": 0.9650328755378723,
+      "grad_norm": 0.9593505263328552,
       "learning_rate": 3.822463768115942e-05,
-      "loss": 0.4269,
+      "loss": 0.4268,
       "step": 650
     },
     {
       "epoch": 0.7065217391304348,
-      "eval_loss": 0.5938716530799866,
-      "eval_runtime": 10.3772,
-      "eval_samples_per_second": 46.255,
+      "eval_loss": 0.5908556580543518,
+      "eval_runtime": 10.3762,
+      "eval_samples_per_second": 46.26,
       "eval_steps_per_second": 2.891,
       "step": 650
     },
     {
       "epoch": 0.717391304347826,
-      "grad_norm": 0.9935194849967957,
+      "grad_norm": 0.9807693362236023,
       "learning_rate": 3.804347826086957e-05,
-      "loss": 0.4223,
+      "loss": 0.4224,
       "step": 660
     },
     {
       "epoch": 0.7282608695652174,
-      "grad_norm": 1.0365718603134155,
+      "grad_norm": 1.0408517122268677,
       "learning_rate": 3.7862318840579716e-05,
       "loss": 0.41,
       "step": 670
     },
     {
       "epoch": 0.7391304347826086,
-      "grad_norm": 0.9572827219963074,
+      "grad_norm": 0.9578678607940674,
       "learning_rate": 3.7681159420289856e-05,
-      "loss": 0.42,
+      "loss": 0.4207,
       "step": 680
     },
     {
       "epoch": 0.75,
-      "grad_norm": 1.034522294998169,
+      "grad_norm": 1.0341449975967407,
       "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.4432,
+      "loss": 0.4435,
       "step": 690
     },
     {
       "epoch": 0.7608695652173914,
-      "grad_norm": 0.9503633379936218,
+      "grad_norm": 0.9478809833526611,
       "learning_rate": 3.7318840579710144e-05,
-      "loss": 0.4464,
+      "loss": 0.4463,
       "step": 700
     },
     {
       "epoch": 0.7608695652173914,
-      "eval_loss": 0.5888856053352356,
-      "eval_runtime": 10.3948,
-      "eval_samples_per_second": 46.177,
-      "eval_steps_per_second": 2.886,
+      "eval_loss": 0.585399866104126,
+      "eval_runtime": 10.3883,
+      "eval_samples_per_second": 46.206,
+      "eval_steps_per_second": 2.888,
       "step": 700
     },
     {
       "epoch": 0.7717391304347826,
-      "grad_norm": 1.038071870803833,
+      "grad_norm": 1.0197566747665405,
       "learning_rate": 3.713768115942029e-05,
       "loss": 0.4234,
       "step": 710
     },
     {
       "epoch": 0.782608695652174,
-      "grad_norm": 1.0283771753311157,
+      "grad_norm": 1.0265061855316162,
       "learning_rate": 3.695652173913043e-05,
-      "loss": 0.4019,
+      "loss": 0.4027,
       "step": 720
     },
     {
       "epoch": 0.7934782608695652,
-      "grad_norm": 0.9672926068305969,
+      "grad_norm": 0.9763724207878113,
       "learning_rate": 3.6775362318840586e-05,
-      "loss": 0.4286,
+      "loss": 0.4284,
       "step": 730
     },
     {
       "epoch": 0.8043478260869565,
-      "grad_norm": 1.0848443508148193,
+      "grad_norm": 1.1108958721160889,
       "learning_rate": 3.6594202898550726e-05,
-      "loss": 0.4196,
+      "loss": 0.4197,
       "step": 740
     },
     {
       "epoch": 0.8152173913043478,
-      "grad_norm": 0.9358295798301697,
+      "grad_norm": 0.938008189201355,
       "learning_rate": 3.641304347826087e-05,
-      "loss": 0.4127,
+      "loss": 0.4133,
       "step": 750
     },
     {
       "epoch": 0.8152173913043478,
-      "eval_loss": 0.5893052220344543,
-      "eval_runtime": 10.3778,
-      "eval_samples_per_second": 46.252,
+      "eval_loss": 0.5860960483551025,
+      "eval_runtime": 10.3772,
+      "eval_samples_per_second": 46.255,
       "eval_steps_per_second": 2.891,
       "step": 750
     },
     {
       "epoch": 0.8260869565217391,
-      "grad_norm": 1.1050411462783813,
+      "grad_norm": 1.106451392173767,
       "learning_rate": 3.6231884057971014e-05,
-      "loss": 0.4335,
+      "loss": 0.4342,
       "step": 760
     },
     {
       "epoch": 0.8369565217391305,
-      "grad_norm": 1.1128437519073486,
+      "grad_norm": 1.1099311113357544,
       "learning_rate": 3.605072463768116e-05,
-      "loss": 0.434,
+      "loss": 0.4345,
       "step": 770
     },
     {
       "epoch": 0.8478260869565217,
-      "grad_norm": 1.0179415941238403,
+      "grad_norm": 1.0185613632202148,
       "learning_rate": 3.58695652173913e-05,
-      "loss": 0.419,
+      "loss": 0.4196,
       "step": 780
     },
     {
       "epoch": 0.8586956521739131,
-      "grad_norm": 0.9764388799667358,
+      "grad_norm": 0.9686940908432007,
       "learning_rate": 3.568840579710145e-05,
-      "loss": 0.4061,
+      "loss": 0.4063,
       "step": 790
     },
     {
       "epoch": 0.8695652173913043,
-      "grad_norm": 0.9722176194190979,
+      "grad_norm": 0.9732295274734497,
       "learning_rate": 3.5507246376811596e-05,
-      "loss": 0.4359,
+      "loss": 0.4365,
       "step": 800
     },
     {
       "epoch": 0.8695652173913043,
-      "eval_loss": 0.5821022391319275,
-      "eval_runtime": 10.3721,
-      "eval_samples_per_second": 46.278,
-      "eval_steps_per_second": 2.892,
+      "eval_loss": 0.5782727003097534,
+      "eval_runtime": 10.4122,
+      "eval_samples_per_second": 46.1,
+      "eval_steps_per_second": 2.881,
       "step": 800
     },
     {
       "epoch": 0.8804347826086957,
-      "grad_norm": 1.1173466444015503,
+      "grad_norm": 1.1204928159713745,
       "learning_rate": 3.532608695652174e-05,
-      "loss": 0.416,
+      "loss": 0.4165,
       "step": 810
     },
     {
       "epoch": 0.8913043478260869,
-      "grad_norm": 1.0256565809249878,
+      "grad_norm": 1.0249269008636475,
       "learning_rate": 3.514492753623188e-05,
-      "loss": 0.4335,
+      "loss": 0.4339,
       "step": 820
     },
     {
       "epoch": 0.9021739130434783,
-      "grad_norm": 0.9599832892417908,
+      "grad_norm": 0.9561118483543396,
       "learning_rate": 3.496376811594203e-05,
-      "loss": 0.4072,
+      "loss": 0.4077,
       "step": 830
     },
     {
       "epoch": 0.9130434782608695,
-      "grad_norm": 1.0554136037826538,
+      "grad_norm": 1.0603339672088623,
       "learning_rate": 3.478260869565218e-05,
-      "loss": 0.4173,
+      "loss": 0.4176,
       "step": 840
     },
     {
       "epoch": 0.9239130434782609,
-      "grad_norm": 0.9580271244049072,
+      "grad_norm": 0.9610790610313416,
       "learning_rate": 3.460144927536232e-05,
-      "loss": 0.418,
+      "loss": 0.4179,
       "step": 850
     },
     {
       "epoch": 0.9239130434782609,
-      "eval_loss": 0.5831677317619324,
-      "eval_runtime": 10.3923,
-      "eval_samples_per_second": 46.188,
-      "eval_steps_per_second": 2.887,
+      "eval_loss": 0.5790796875953674,
+      "eval_runtime": 10.4182,
+      "eval_samples_per_second": 46.073,
+      "eval_steps_per_second": 2.88,
       "step": 850
     },
     {
       "epoch": 0.9347826086956522,
-      "grad_norm": 1.0673909187316895,
+      "grad_norm": 1.0670075416564941,
       "learning_rate": 3.4420289855072465e-05,
-      "loss": 0.4115,
+      "loss": 0.4114,
       "step": 860
     },
     {
       "epoch": 0.9456521739130435,
-      "grad_norm": 1.0771151781082153,
+      "grad_norm": 1.070837140083313,
       "learning_rate": 3.423913043478261e-05,
-      "loss": 0.4038,
+      "loss": 0.4034,
       "step": 870
     },
     {
       "epoch": 0.9565217391304348,
-      "grad_norm": 1.0083266496658325,
+      "grad_norm": 1.0078990459442139,
       "learning_rate": 3.405797101449276e-05,
-      "loss": 0.399,
+      "loss": 0.3985,
       "step": 880
     },
     {
       "epoch": 0.967391304347826,
-      "grad_norm": 1.0740772485733032,
+      "grad_norm": 1.0645533800125122,
       "learning_rate": 3.38768115942029e-05,
-      "loss": 0.4231,
+      "loss": 0.4238,
       "step": 890
     },
     {
       "epoch": 0.9782608695652174,
-      "grad_norm": 0.9597785472869873,
+      "grad_norm": 0.9449084401130676,
       "learning_rate": 3.369565217391305e-05,
-      "loss": 0.4151,
+      "loss": 0.4146,
       "step": 900
     },
     {
       "epoch": 0.9782608695652174,
-      "eval_loss": 0.5803946852684021,
-      "eval_runtime": 10.3767,
-      "eval_samples_per_second": 46.258,
-      "eval_steps_per_second": 2.891,
+      "eval_loss": 0.5765789151191711,
+      "eval_runtime": 10.4319,
+      "eval_samples_per_second": 46.013,
+      "eval_steps_per_second": 2.876,
       "step": 900
     },
     {
       "epoch": 0.9891304347826086,
-      "grad_norm": 1.1659215688705444,
+      "grad_norm": 1.1554139852523804,
       "learning_rate": 3.351449275362319e-05,
-      "loss": 0.397,
+      "loss": 0.3972,
       "step": 910
     },
     {
       "epoch": 1.0,
-      "grad_norm": 2.251577138900757,
+      "grad_norm": 2.253143787384033,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.4355,
+      "loss": 0.4359,
       "step": 920
     },
     {
       "epoch": 1.0108695652173914,
-      "grad_norm": 1.1286661624908447,
+      "grad_norm": 1.124770164489746,
       "learning_rate": 3.3152173913043475e-05,
-      "loss": 0.3536,
+      "loss": 0.3533,
       "step": 930
     },
     {
       "epoch": 1.0217391304347827,
-      "grad_norm": 1.1189433336257935,
+      "grad_norm": 1.1178290843963623,
       "learning_rate": 3.297101449275363e-05,
-      "loss": 0.3596,
+      "loss": 0.3599,
       "step": 940
     },
     {
       "epoch": 1.0326086956521738,
-      "grad_norm": 1.0220948457717896,
+      "grad_norm": 1.0256346464157104,
       "learning_rate": 3.278985507246377e-05,
       "loss": 0.3632,
       "step": 950
     },
     {
       "epoch": 1.0326086956521738,
-      "eval_loss": 0.5912913680076599,
-      "eval_runtime": 10.3768,
-      "eval_samples_per_second": 46.257,
-      "eval_steps_per_second": 2.891,
+      "eval_loss": 0.5881566405296326,
+      "eval_runtime": 10.4627,
+      "eval_samples_per_second": 45.877,
+      "eval_steps_per_second": 2.867,
       "step": 950
     },
     {
       "epoch": 1.0434782608695652,
-      "grad_norm": 1.0051358938217163,
+      "grad_norm": 1.0145591497421265,
       "learning_rate": 3.260869565217392e-05,
-      "loss": 0.3648,
+      "loss": 0.3647,
       "step": 960
     },
     {
       "epoch": 1.0543478260869565,
-      "grad_norm": 1.1821051836013794,
+      "grad_norm": 1.2014787197113037,
       "learning_rate": 3.242753623188406e-05,
-      "loss": 0.3705,
+      "loss": 0.37,
       "step": 970
     },
     {
       "epoch": 1.065217391304348,
-      "grad_norm": 1.1006765365600586,
+      "grad_norm": 1.0973201990127563,
       "learning_rate": 3.2246376811594205e-05,
-      "loss": 0.3606,
+      "loss": 0.3608,
       "step": 980
     },
     {
       "epoch": 1.0760869565217392,
-      "grad_norm": 1.1193490028381348,
+      "grad_norm": 1.120807409286499,
       "learning_rate": 3.2065217391304345e-05,
-      "loss": 0.3471,
+      "loss": 0.3466,
       "step": 990
     },
     {
       "epoch": 1.0869565217391304,
-      "grad_norm": 1.1935309171676636,
+      "grad_norm": 1.1913529634475708,
       "learning_rate": 3.188405797101449e-05,
-      "loss": 0.3667,
+      "loss": 0.3671,
       "step": 1000
     },
     {
       "epoch": 1.0869565217391304,
-      "eval_loss": 0.5953035950660706,
-      "eval_runtime": 10.3685,
-      "eval_samples_per_second": 46.294,
-      "eval_steps_per_second": 2.893,
+      "eval_loss": 0.591897189617157,
+      "eval_runtime": 10.4218,
+      "eval_samples_per_second": 46.057,
+      "eval_steps_per_second": 2.879,
       "step": 1000
     }
   ],

checkpoint-1000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:40ebe6ec892e02f02c5df3d2f37d3f9302b8a2381547d3bdf4d52f5e045e3cdc
+oid sha256:a3712c5fdd6e5241e8ea26ae62b28612fc514ec88e8a03bc6c1b3a46acfc611a
 size 5816