Upload training_metrics.json with huggingface_hub

training_metrics.json

@@ -3,377 +3,377 @@
   "dataset": "Nanthasit/sakthai-combined-v5",
   "log": [
     {
-      "loss": 2.068629264831543,
-      "grad_norm": 1.109375,
+      "loss": 2.06823673248291,
+      "grad_norm": 1.140625,
       "learning_rate": 3e-05,
-      "entropy": 0.8546215698122979,
+      "entropy": 0.8546648569405079,
       "num_tokens": 101440.0,
-      "mean_token_accuracy": 0.6594636622816324,
+      "mean_token_accuracy": 0.6593547906726599,
       "epoch": 0.26229508196721313,
       "step": 10
     },
     {
-      "loss": 1.8951040267944337,
-      "grad_norm": 0.69140625,
+      "loss": 1.894563865661621,
+      "grad_norm": 0.69921875,
       "learning_rate": 4.9298245614035086e-05,
-      "entropy": 0.9977125555276871,
+      "entropy": 0.9990503549575805,
       "num_tokens": 203698.0,
-      "mean_token_accuracy": 0.6644318636506796,
+      "mean_token_accuracy": 0.6645173218101263,
       "epoch": 0.5245901639344263,
       "step": 20
     },
     {
-      "loss": 1.543001365661621,
-      "grad_norm": 0.359375,
+      "loss": 1.5417596817016601,
+      "grad_norm": 0.361328125,
       "learning_rate": 4.754385964912281e-05,
-      "entropy": 1.2034106969833374,
+      "entropy": 1.209834586828947,
       "num_tokens": 304484.0,
-      "mean_token_accuracy": 0.6810707345604896,
+      "mean_token_accuracy": 0.6815106488764286,
       "epoch": 0.7868852459016393,
       "step": 30
     },
     {
-      "loss": 1.4598698616027832,
-      "grad_norm": 0.419921875,
+      "loss": 1.454429817199707,
+      "grad_norm": 0.423828125,
       "learning_rate": 4.5789473684210527e-05,
-      "entropy": 1.3983221878744152,
+      "entropy": 1.4112283179204759,
       "num_tokens": 397014.0,
-      "mean_token_accuracy": 0.6836717132836172,
+      "mean_token_accuracy": 0.684118391308066,
       "epoch": 1.0262295081967212,
       "step": 40
     },
     {
-      "loss": 1.2653812408447265,
-      "grad_norm": 0.41796875,
+      "loss": 1.2588438987731934,
+      "grad_norm": 0.416015625,
       "learning_rate": 4.403508771929824e-05,
-      "entropy": 1.3433212354779243,
+      "entropy": 1.355584604293108,
       "num_tokens": 497195.0,
-      "mean_token_accuracy": 0.7267363253980875,
+      "mean_token_accuracy": 0.7292849004268647,
       "epoch": 1.2885245901639344,
       "step": 50
     },
     {
-      "eval_loss": 0.9654702544212341,
-      "eval_runtime": 17.0259,
-      "eval_samples_per_second": 2.937,
+      "eval_loss": 0.9597824811935425,
+      "eval_runtime": 17.0145,
+      "eval_samples_per_second": 2.939,
       "eval_steps_per_second": 0.411,
-      "eval_entropy": 1.0037876112120492,
+      "eval_entropy": 1.0092762282916479,
       "eval_num_tokens": 497195.0,
-      "eval_mean_token_accuracy": 0.7851139051573617,
+      "eval_mean_token_accuracy": 0.7874619875635419,
       "epoch": 1.2885245901639344,
       "step": 50
     },
     {
-      "loss": 0.9160248756408691,
-      "grad_norm": 0.77734375,
+      "loss": 0.9087855339050293,
+      "grad_norm": 0.765625,
       "learning_rate": 4.228070175438597e-05,
-      "entropy": 0.9984242603182792,
+      "entropy": 1.0016439206898213,
       "num_tokens": 599800.0,
-      "mean_token_accuracy": 0.7878576569259167,
+      "mean_token_accuracy": 0.7895643316209316,
       "epoch": 1.5508196721311476,
       "step": 60
     },
     {
-      "loss": 0.8177807807922364,
-      "grad_norm": 0.384765625,
+      "loss": 0.8143481254577637,
+      "grad_norm": 0.3671875,
       "learning_rate": 4.0526315789473684e-05,
-      "entropy": 0.794848458096385,
+      "entropy": 0.7972811751067639,
       "num_tokens": 701131.0,
-      "mean_token_accuracy": 0.8004833266139031,
+      "mean_token_accuracy": 0.8020147792994976,
       "epoch": 1.8131147540983608,
       "step": 70
     },
     {
-      "loss": 0.697277307510376,
-      "grad_norm": 0.2451171875,
+      "loss": 0.6968278408050537,
+      "grad_norm": 0.2470703125,
       "learning_rate": 3.877192982456141e-05,
-      "entropy": 0.693726461635877,
+      "entropy": 0.6991581267690006,
       "num_tokens": 795040.0,
-      "mean_token_accuracy": 0.8268714155236335,
+      "mean_token_accuracy": 0.8275373725041951,
       "epoch": 2.0524590163934424,
       "step": 80
     },
     {
-      "loss": 0.7282017707824707,
-      "grad_norm": 0.2734375,
+      "loss": 0.7272923469543457,
+      "grad_norm": 0.265625,
       "learning_rate": 3.7017543859649124e-05,
-      "entropy": 0.692317858338356,
+      "entropy": 0.6977068085223437,
       "num_tokens": 901989.0,
-      "mean_token_accuracy": 0.8255993157625199,
+      "mean_token_accuracy": 0.826152490079403,
       "epoch": 2.314754098360656,
       "step": 90
     },
     {
-      "loss": 0.6321603775024414,
-      "grad_norm": 0.298828125,
+      "loss": 0.6305916786193848,
+      "grad_norm": 0.26171875,
       "learning_rate": 3.526315789473684e-05,
-      "entropy": 0.6189962957054377,
+      "entropy": 0.6202053360641002,
       "num_tokens": 1006326.0,
-      "mean_token_accuracy": 0.8391064219176769,
+      "mean_token_accuracy": 0.8399624571204185,
       "epoch": 2.577049180327869,
       "step": 100
     },
     {
-      "eval_loss": 0.5240724682807922,
-      "eval_runtime": 17.0105,
-      "eval_samples_per_second": 2.939,
-      "eval_steps_per_second": 0.412,
-      "eval_entropy": 0.528444230556488,
+      "eval_loss": 0.5223863124847412,
+      "eval_runtime": 17.0185,
+      "eval_samples_per_second": 2.938,
+      "eval_steps_per_second": 0.411,
+      "eval_entropy": 0.5274915056569236,
       "eval_num_tokens": 1006326.0,
-      "eval_mean_token_accuracy": 0.8560459698949542,
+      "eval_mean_token_accuracy": 0.8570880378995623,
       "epoch": 2.577049180327869,
       "step": 100
     },
     {
-      "loss": 0.6537245750427246,
-      "grad_norm": 0.2470703125,
+      "loss": 0.6520613670349121,
+      "grad_norm": 0.259765625,
       "learning_rate": 3.3508771929824564e-05,
-      "entropy": 0.5889126423746347,
+      "entropy": 0.5896702250465751,
       "num_tokens": 1101001.0,
-      "mean_token_accuracy": 0.8503337949514389,
+      "mean_token_accuracy": 0.8507260680198669,
       "epoch": 2.839344262295082,
       "step": 110
     },
     {
-      "loss": 0.7455410480499267,
-      "grad_norm": 0.251953125,
+      "loss": 0.7440696239471436,
+      "grad_norm": 0.265625,
       "learning_rate": 3.175438596491228e-05,
-      "entropy": 0.6424324704359655,
+      "entropy": 0.6408274096168883,
       "num_tokens": 1192903.0,
-      "mean_token_accuracy": 0.8340811778421271,
+      "mean_token_accuracy": 0.8347789571709829,
       "epoch": 3.078688524590164,
       "step": 120
     },
     {
-      "loss": 0.6715424537658692,
-      "grad_norm": 0.2236328125,
+      "loss": 0.6696242332458496,
+      "grad_norm": 0.2138671875,
       "learning_rate": 3e-05,
-      "entropy": 0.6274289535358548,
+      "entropy": 0.6266658639535307,
       "num_tokens": 1295698.0,
-      "mean_token_accuracy": 0.8365047372877598,
+      "mean_token_accuracy": 0.836974135786295,
       "epoch": 3.3409836065573773,
       "step": 130
     },
     {
-      "loss": 0.6283944129943848,
-      "grad_norm": 0.318359375,
+      "loss": 0.6262566089630127,
+      "grad_norm": 0.322265625,
       "learning_rate": 2.824561403508772e-05,
-      "entropy": 0.622589353658259,
+      "entropy": 0.6221931139007211,
       "num_tokens": 1396196.0,
-      "mean_token_accuracy": 0.8458590045571327,
+      "mean_token_accuracy": 0.8461123399436474,
       "epoch": 3.60327868852459,
       "step": 140
     },
     {
-      "loss": 0.5661543846130371,
-      "grad_norm": 0.318359375,
+      "loss": 0.5626291751861572,
+      "grad_norm": 0.314453125,
       "learning_rate": 2.6491228070175438e-05,
-      "entropy": 0.5624578293412924,
+      "entropy": 0.5584697874262929,
       "num_tokens": 1494364.0,
-      "mean_token_accuracy": 0.8553425297141075,
+      "mean_token_accuracy": 0.8564164794981479,
       "epoch": 3.865573770491803,
       "step": 150
     },
     {
-      "eval_loss": 0.4366666078567505,
-      "eval_runtime": 17.0114,
+      "eval_loss": 0.430507630109787,
+      "eval_runtime": 17.0122,
       "eval_samples_per_second": 2.939,
       "eval_steps_per_second": 0.411,
-      "eval_entropy": 0.4516720644065312,
+      "eval_entropy": 0.4446746025766645,
       "eval_num_tokens": 1494364.0,
-      "eval_mean_token_accuracy": 0.8738486085619245,
+      "eval_mean_token_accuracy": 0.8762749433517456,
       "epoch": 3.865573770491803,
       "step": 150
     },
     {
-      "loss": 0.5499395370483399,
-      "grad_norm": 0.30078125,
+      "loss": 0.5449130535125732,
+      "grad_norm": 0.275390625,
       "learning_rate": 2.4736842105263158e-05,
-      "entropy": 0.5944871929002135,
+      "entropy": 0.5905464154808489,
       "num_tokens": 1586359.0,
-      "mean_token_accuracy": 0.8487225819940436,
+      "mean_token_accuracy": 0.8499162156287938,
       "epoch": 4.104918032786885,
       "step": 160
     },
     {
-      "loss": 0.5848193645477295,
-      "grad_norm": 0.287109375,
+      "loss": 0.5826212882995605,
+      "grad_norm": 0.2734375,
       "learning_rate": 2.2982456140350878e-05,
-      "entropy": 0.5599356807768345,
+      "entropy": 0.5561527205631137,
       "num_tokens": 1685436.0,
-      "mean_token_accuracy": 0.8498164191842079,
+      "mean_token_accuracy": 0.8504404634237289,
       "epoch": 4.367213114754098,
       "step": 170
     },
     {
-      "loss": 0.5726899623870849,
-      "grad_norm": 0.296875,
+      "loss": 0.5708770275115966,
+      "grad_norm": 0.287109375,
       "learning_rate": 2.1228070175438598e-05,
-      "entropy": 0.547566929180175,
+      "entropy": 0.5452653610147535,
       "num_tokens": 1786379.0,
-      "mean_token_accuracy": 0.8605796664953231,
+      "mean_token_accuracy": 0.86077646240592,
       "epoch": 4.629508196721312,
       "step": 180
     },
     {
-      "loss": 0.594398021697998,
+      "loss": 0.5923133373260498,
       "grad_norm": 0.248046875,
       "learning_rate": 1.9473684210526315e-05,
-      "entropy": 0.5992783515714108,
+      "entropy": 0.5972572009079158,
       "num_tokens": 1892366.0,
-      "mean_token_accuracy": 0.8466935507953167,
+      "mean_token_accuracy": 0.8472615011036396,
       "epoch": 4.891803278688524,
       "step": 190
     },
     {
-      "loss": 0.7168087005615235,
-      "grad_norm": 0.28515625,
+      "loss": 0.7141714096069336,
+      "grad_norm": 0.298828125,
       "learning_rate": 1.7719298245614035e-05,
-      "entropy": 0.595610975504738,
+      "entropy": 0.5953059407742056,
       "num_tokens": 1984266.0,
-      "mean_token_accuracy": 0.8466557786889273,
+      "mean_token_accuracy": 0.8461490273475647,
       "epoch": 5.131147540983607,
       "step": 200
     },
     {
-      "eval_loss": 0.39171460270881653,
-      "eval_runtime": 17.0117,
-      "eval_samples_per_second": 2.939,
-      "eval_steps_per_second": 0.411,
-      "eval_entropy": 0.39902772647993906,
+      "eval_loss": 0.3904975950717926,
+      "eval_runtime": 17.0097,
+      "eval_samples_per_second": 2.94,
+      "eval_steps_per_second": 0.412,
+      "eval_entropy": 0.39667304924556185,
       "eval_num_tokens": 1984266.0,
-      "eval_mean_token_accuracy": 0.8864545055798122,
+      "eval_mean_token_accuracy": 0.8853033610752651,
       "epoch": 5.131147540983607,
       "step": 200
     },
     {
-      "loss": 0.5475130558013916,
-      "grad_norm": 0.279296875,
+      "loss": 0.5458156108856201,
+      "grad_norm": 0.267578125,
       "learning_rate": 1.5964912280701755e-05,
-      "entropy": 0.5358397019095719,
+      "entropy": 0.5346348247490823,
       "num_tokens": 2086788.0,
-      "mean_token_accuracy": 0.8598736800253391,
+      "mean_token_accuracy": 0.8602009244263172,
       "epoch": 5.39344262295082,
       "step": 210
     },
     {
-      "loss": 0.5264423847198486,
-      "grad_norm": 0.283203125,
+      "loss": 0.5247596263885498,
+      "grad_norm": 0.25390625,
       "learning_rate": 1.4210526315789475e-05,
-      "entropy": 0.49262598659843204,
+      "entropy": 0.4917796904221177,
       "num_tokens": 2188133.0,
-      "mean_token_accuracy": 0.8688649766147136,
+      "mean_token_accuracy": 0.8693851955235005,
       "epoch": 5.655737704918033,
       "step": 220
     },
     {
-      "loss": 0.6182221889495849,
-      "grad_norm": 0.298828125,
+      "loss": 0.6164801597595215,
+      "grad_norm": 0.259765625,
       "learning_rate": 1.2456140350877193e-05,
-      "entropy": 0.5918203789740801,
+      "entropy": 0.5906232800334692,
       "num_tokens": 2289867.0,
-      "mean_token_accuracy": 0.8466856233775616,
+      "mean_token_accuracy": 0.8469160348176956,
       "epoch": 5.918032786885246,
       "step": 230
     },
     {
-      "loss": 0.5268895626068115,
-      "grad_norm": 0.203125,
+      "loss": 0.5254185676574707,
+      "grad_norm": 0.2021484375,
       "learning_rate": 1.0701754385964913e-05,
-      "entropy": 0.5374027302310075,
+      "entropy": 0.534850709038238,
       "num_tokens": 2384185.0,
-      "mean_token_accuracy": 0.857935513535591,
+      "mean_token_accuracy": 0.8580893457752384,
       "epoch": 6.157377049180328,
       "step": 240
     },
     {
-      "loss": 0.5403360843658447,
-      "grad_norm": 0.2255859375,
+      "loss": 0.5394423007965088,
+      "grad_norm": 0.2333984375,
       "learning_rate": 8.947368421052632e-06,
-      "entropy": 0.5230959854088724,
+      "entropy": 0.5212304137647152,
       "num_tokens": 2482981.0,
-      "mean_token_accuracy": 0.8593067392706871,
+      "mean_token_accuracy": 0.8591319762170315,
       "epoch": 6.419672131147541,
       "step": 250
     },
     {
-      "eval_loss": 0.38528522849082947,
-      "eval_runtime": 17.0169,
-      "eval_samples_per_second": 2.938,
+      "eval_loss": 0.38513797521591187,
+      "eval_runtime": 17.0114,
+      "eval_samples_per_second": 2.939,
       "eval_steps_per_second": 0.411,
-      "eval_entropy": 0.3912909414087023,
+      "eval_entropy": 0.38969274929591585,
       "eval_num_tokens": 2482981.0,
-      "eval_mean_token_accuracy": 0.8874894210270473,
+      "eval_mean_token_accuracy": 0.8868268047060285,
       "epoch": 6.419672131147541,
       "step": 250
     },
     {
-      "loss": 0.5694822311401367,
-      "grad_norm": 0.26171875,
+      "loss": 0.5682054042816163,
+      "grad_norm": 0.25390625,
       "learning_rate": 7.192982456140351e-06,
-      "entropy": 0.5577683749608695,
+      "entropy": 0.5569836483336985,
       "num_tokens": 2584342.0,
-      "mean_token_accuracy": 0.8576721116900444,
+      "mean_token_accuracy": 0.8579698376357555,
       "epoch": 6.6819672131147545,
       "step": 260
     },
     {
-      "loss": 0.5874075889587402,
-      "grad_norm": 0.390625,
+      "loss": 0.5858804702758789,
+      "grad_norm": 0.357421875,
       "learning_rate": 5.43859649122807e-06,
-      "entropy": 0.5574779569171369,
+      "entropy": 0.5571649130433798,
       "num_tokens": 2688295.0,
-      "mean_token_accuracy": 0.8561453945934773,
+      "mean_token_accuracy": 0.856262207776308,
       "epoch": 6.944262295081967,
       "step": 270
     },
     {
-      "loss": 0.6308839321136475,
-      "grad_norm": 0.26171875,
+      "loss": 0.6284305572509765,
+      "grad_norm": 0.251953125,
       "learning_rate": 3.6842105263157892e-06,
-      "entropy": 0.5019811069312161,
+      "entropy": 0.5009839224488768,
       "num_tokens": 2781223.0,
-      "mean_token_accuracy": 0.8665594115649184,
+      "mean_token_accuracy": 0.8666133962265433,
       "epoch": 7.18360655737705,
       "step": 280
     },
     {
-      "loss": 0.6411924839019776,
-      "grad_norm": 0.2236328125,
+      "loss": 0.6396284103393555,
+      "grad_norm": 0.228515625,
       "learning_rate": 1.929824561403509e-06,
-      "entropy": 0.6177210337482393,
+      "entropy": 0.6168746447190643,
       "num_tokens": 2883509.0,
-      "mean_token_accuracy": 0.8411735735833645,
+      "mean_token_accuracy": 0.8415784001350403,
       "epoch": 7.445901639344262,
       "step": 290
     },
     {
-      "loss": 0.5143293380737305,
-      "grad_norm": 0.287109375,
+      "loss": 0.5129966735839844,
+      "grad_norm": 0.28515625,
       "learning_rate": 1.7543859649122808e-07,
-      "entropy": 0.506786679290235,
+      "entropy": 0.5056241630576551,
       "num_tokens": 2979817.0,
-      "mean_token_accuracy": 0.8670562252402305,
+      "mean_token_accuracy": 0.867512446641922,
       "epoch": 7.7081967213114755,
       "step": 300
     },
     {
-      "eval_loss": 0.3845280110836029,
-      "eval_runtime": 17.0166,
+      "eval_loss": 0.38443076610565186,
+      "eval_runtime": 17.0197,
       "eval_samples_per_second": 2.938,
       "eval_steps_per_second": 0.411,
-      "eval_entropy": 0.3884585755211966,
+      "eval_entropy": 0.3875300394637244,
       "eval_num_tokens": 2979817.0,
-      "eval_mean_token_accuracy": 0.8877047896385193,
+      "eval_mean_token_accuracy": 0.8867072548185077,
       "epoch": 7.7081967213114755,
       "step": 300
     },
     {
-      "train_runtime": 4734.1897,
+      "train_runtime": 4733.1969,
       "train_samples_per_second": 1.014,
       "train_steps_per_second": 0.063,
       "total_flos": 1.5760695558322176e+17,
-      "train_loss": 0.8003380727767945,
+      "train_loss": 0.7980758142471314,
       "epoch": 7.7081967213114755,
       "step": 300
     }