Fin de l'entraînement

README.md

@@ -14,9 +14,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0834
-- Mse: 0.1351
-- Mae: 0.3171
+- Loss: 0.0859
+- Mse: 0.1291
+- Mae: 0.3120
 
 ## Model description
 

all_results.json

@@ -1,14 +1,14 @@
 {
-    "epoch": 50.0,
-    "eval_loss": 0.05992668867111206,
-    "eval_mae": 0.3671800494194031,
-    "eval_mse": 0.16954755783081055,
-    "eval_runtime": 2.0879,
-    "eval_samples_per_second": 4.789,
-    "eval_steps_per_second": 0.958,
+    "epoch": 18.0,
+    "eval_loss": 0.08588916808366776,
+    "eval_mae": 0.3119707405567169,
+    "eval_mse": 0.1291341334581375,
+    "eval_runtime": 9.945,
+    "eval_samples_per_second": 7.642,
+    "eval_steps_per_second": 1.006,
     "total_flos": 0.0,
-    "train_loss": 0.08142914772033691,
-    "train_runtime": 429.9896,
-    "train_samples_per_second": 9.07,
-    "train_steps_per_second": 0.233
+    "train_loss": 0.2600840449333191,
+    "train_runtime": 1002.0455,
+    "train_samples_per_second": 6.048,
+    "train_steps_per_second": 0.18
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 50.0,
-    "eval_loss": 0.05992668867111206,
-    "eval_mae": 0.3671800494194031,
-    "eval_mse": 0.16954755783081055,
-    "eval_runtime": 2.0879,
-    "eval_samples_per_second": 4.789,
-    "eval_steps_per_second": 0.958
+    "epoch": 18.0,
+    "eval_loss": 0.08588916808366776,
+    "eval_mae": 0.3119707405567169,
+    "eval_mse": 0.1291341334581375,
+    "eval_runtime": 9.945,
+    "eval_samples_per_second": 7.642,
+    "eval_steps_per_second": 1.006
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
+    "epoch": 18.0,
     "total_flos": 0.0,
-    "train_loss": 0.08142914772033691,
-    "train_runtime": 429.9896,
-    "train_samples_per_second": 9.07,
-    "train_steps_per_second": 0.233
+    "train_loss": 0.2600840449333191,
+    "train_runtime": 1002.0455,
+    "train_samples_per_second": 6.048,
+    "train_steps_per_second": 0.18
 }

trainer_state.json

@@ -1,542 +1,236 @@
 {
-  "best_global_step": 58,
-  "best_metric": 0.16954755783081055,
-  "best_model_checkpoint": "./vit_focus/checkpoint-58",
-  "epoch": 50.0,
+  "best_global_step": 100,
+  "best_metric": 0.1291341334581375,
+  "best_model_checkpoint": "./vit_focus/checkpoint-100",
+  "epoch": 18.0,
   "eval_steps": 500,
-  "global_step": 100,
+  "global_step": 180,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_loss": 0.12364959716796875,
-      "eval_mae": 0.41142117977142334,
-      "eval_mse": 0.2097923755645752,
-      "eval_runtime": 1.3273,
-      "eval_samples_per_second": 7.534,
-      "eval_steps_per_second": 1.507,
-      "step": 2
+      "eval_loss": 0.11766688525676727,
+      "eval_mae": 0.3475768268108368,
+      "eval_mse": 0.1613956242799759,
+      "eval_runtime": 10.0195,
+      "eval_samples_per_second": 7.585,
+      "eval_steps_per_second": 0.998,
+      "step": 10
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.1196303591132164,
-      "eval_mae": 0.40995553135871887,
-      "eval_mse": 0.2083383947610855,
-      "eval_runtime": 1.2492,
-      "eval_samples_per_second": 8.005,
-      "eval_steps_per_second": 1.601,
-      "step": 4
+      "eval_loss": 0.09316065162420273,
+      "eval_mae": 0.32877787947654724,
+      "eval_mse": 0.14106006920337677,
+      "eval_runtime": 9.4019,
+      "eval_samples_per_second": 8.083,
+      "eval_steps_per_second": 1.064,
+      "step": 20
     },
     {
       "epoch": 3.0,
-      "eval_loss": 0.11328934133052826,
-      "eval_mae": 0.4075067639350891,
-      "eval_mse": 0.20581738650798798,
-      "eval_runtime": 1.1974,
-      "eval_samples_per_second": 8.352,
-      "eval_steps_per_second": 1.67,
-      "step": 6
+      "eval_loss": 0.09138184040784836,
+      "eval_mae": 0.3351140022277832,
+      "eval_mse": 0.14778681099414825,
+      "eval_runtime": 9.4168,
+      "eval_samples_per_second": 8.071,
+      "eval_steps_per_second": 1.062,
+      "step": 30
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 3.422793388366699,
+      "learning_rate": 7.833333333333333e-05,
+      "loss": 0.3379,
+      "step": 40
     },
     {
       "epoch": 4.0,
-      "eval_loss": 0.10604934394359589,
-      "eval_mae": 0.4042162597179413,
-      "eval_mse": 0.20277394354343414,
-      "eval_runtime": 1.202,
-      "eval_samples_per_second": 8.32,
-      "eval_steps_per_second": 1.664,
-      "step": 8
+      "eval_loss": 0.07990685850381851,
+      "eval_mae": 0.3216111361980438,
+      "eval_mse": 0.1368674337863922,
+      "eval_runtime": 9.6691,
+      "eval_samples_per_second": 7.86,
+      "eval_steps_per_second": 1.034,
+      "step": 40
     },
     {
       "epoch": 5.0,
-      "eval_loss": 0.10327418148517609,
-      "eval_mae": 0.4024171531200409,
-      "eval_mse": 0.20201840996742249,
-      "eval_runtime": 1.3089,
-      "eval_samples_per_second": 7.64,
-      "eval_steps_per_second": 1.528,
-      "step": 10
+      "eval_loss": 0.07730501145124435,
+      "eval_mae": 0.3210676312446594,
+      "eval_mse": 0.1364751160144806,
+      "eval_runtime": 9.6377,
+      "eval_samples_per_second": 7.886,
+      "eval_steps_per_second": 1.038,
+      "step": 50
     },
     {
       "epoch": 6.0,
-      "eval_loss": 0.09384196996688843,
-      "eval_mae": 0.3968161642551422,
-      "eval_mse": 0.19719576835632324,
-      "eval_runtime": 1.2505,
-      "eval_samples_per_second": 7.997,
-      "eval_steps_per_second": 1.599,
-      "step": 12
+      "eval_loss": 0.08424032479524612,
+      "eval_mae": 0.3250683844089508,
+      "eval_mse": 0.14131243526935577,
+      "eval_runtime": 9.4489,
+      "eval_samples_per_second": 8.043,
+      "eval_steps_per_second": 1.058,
+      "step": 60
     },
     {
       "epoch": 7.0,
-      "eval_loss": 0.10823704302310944,
-      "eval_mae": 0.40178966522216797,
-      "eval_mse": 0.2021051049232483,
-      "eval_runtime": 1.2484,
-      "eval_samples_per_second": 8.01,
-      "eval_steps_per_second": 1.602,
-      "step": 14
+      "eval_loss": 0.08766830712556839,
+      "eval_mae": 0.3169863820075989,
+      "eval_mse": 0.13288582861423492,
+      "eval_runtime": 10.5663,
+      "eval_samples_per_second": 7.193,
+      "eval_steps_per_second": 0.946,
+      "step": 70
     },
     {
       "epoch": 8.0,
-      "eval_loss": 0.0798194482922554,
-      "eval_mae": 0.3889862298965454,
-      "eval_mse": 0.18965376913547516,
-      "eval_runtime": 1.2049,
-      "eval_samples_per_second": 8.3,
-      "eval_steps_per_second": 1.66,
-      "step": 16
+      "grad_norm": 3.210268020629883,
+      "learning_rate": 5.6111111111111114e-05,
+      "loss": 0.2829,
+      "step": 80
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.08300431817770004,
+      "eval_mae": 0.3192159831523895,
+      "eval_mse": 0.1370069533586502,
+      "eval_runtime": 9.75,
+      "eval_samples_per_second": 7.795,
+      "eval_steps_per_second": 1.026,
+      "step": 80
     },
     {
       "epoch": 9.0,
-      "eval_loss": 0.10108314454555511,
-      "eval_mae": 0.39635446667671204,
-      "eval_mse": 0.19831416010856628,
-      "eval_runtime": 1.2689,
-      "eval_samples_per_second": 7.881,
-      "eval_steps_per_second": 1.576,
-      "step": 18
+      "eval_loss": 0.08274427056312561,
+      "eval_mae": 0.32542383670806885,
+      "eval_mse": 0.140847310423851,
+      "eval_runtime": 9.5472,
+      "eval_samples_per_second": 7.96,
+      "eval_steps_per_second": 1.047,
+      "step": 90
     },
     {
       "epoch": 10.0,
-      "eval_loss": 0.08742986619472504,
-      "eval_mae": 0.3887980282306671,
-      "eval_mse": 0.19129884243011475,
-      "eval_runtime": 1.2793,
-      "eval_samples_per_second": 7.817,
-      "eval_steps_per_second": 1.563,
-      "step": 20
+      "eval_loss": 0.08588916808366776,
+      "eval_mae": 0.3119707405567169,
+      "eval_mse": 0.1291341334581375,
+      "eval_runtime": 10.6888,
+      "eval_samples_per_second": 7.11,
+      "eval_steps_per_second": 0.936,
+      "step": 100
     },
     {
       "epoch": 11.0,
-      "eval_loss": 0.06867878139019012,
-      "eval_mae": 0.3799101710319519,
-      "eval_mse": 0.18114082515239716,
-      "eval_runtime": 1.3093,
-      "eval_samples_per_second": 7.637,
-      "eval_steps_per_second": 1.527,
-      "step": 22
+      "eval_loss": 0.10085263103246689,
+      "eval_mae": 0.330853134393692,
+      "eval_mse": 0.14782297611236572,
+      "eval_runtime": 9.484,
+      "eval_samples_per_second": 8.013,
+      "eval_steps_per_second": 1.054,
+      "step": 110
     },
     {
       "epoch": 12.0,
-      "eval_loss": 0.08115904778242111,
-      "eval_mae": 0.38764962553977966,
-      "eval_mse": 0.18922746181488037,
-      "eval_runtime": 1.1917,
-      "eval_samples_per_second": 8.392,
-      "eval_steps_per_second": 1.678,
-      "step": 24
+      "grad_norm": 2.0312862396240234,
+      "learning_rate": 3.388888888888889e-05,
+      "loss": 0.2461,
+      "step": 120
+    },
+    {
+      "epoch": 12.0,
+      "eval_loss": 0.08102226257324219,
+      "eval_mae": 0.3174217641353607,
+      "eval_mse": 0.135166734457016,
+      "eval_runtime": 9.5097,
+      "eval_samples_per_second": 7.992,
+      "eval_steps_per_second": 1.052,
+      "step": 120
     },
     {
       "epoch": 13.0,
-      "eval_loss": 0.07042384147644043,
-      "eval_mae": 0.3812001347541809,
-      "eval_mse": 0.1831439584493637,
-      "eval_runtime": 1.2731,
-      "eval_samples_per_second": 7.855,
-      "eval_steps_per_second": 1.571,
-      "step": 26
+      "eval_loss": 0.08370836824178696,
+      "eval_mae": 0.31886667013168335,
+      "eval_mse": 0.13684500753879547,
+      "eval_runtime": 9.4388,
+      "eval_samples_per_second": 8.052,
+      "eval_steps_per_second": 1.059,
+      "step": 130
     },
     {
       "epoch": 14.0,
-      "eval_loss": 0.057908277958631516,
-      "eval_mae": 0.37130409479141235,
-      "eval_mse": 0.17441079020500183,
-      "eval_runtime": 3.0777,
-      "eval_samples_per_second": 3.249,
-      "eval_steps_per_second": 0.65,
-      "step": 28
+      "eval_loss": 0.08873885869979858,
+      "eval_mae": 0.32301220297813416,
+      "eval_mse": 0.1409922093153,
+      "eval_runtime": 9.629,
+      "eval_samples_per_second": 7.893,
+      "eval_steps_per_second": 1.039,
+      "step": 140
     },
     {
       "epoch": 15.0,
-      "eval_loss": 0.06775574386119843,
-      "eval_mae": 0.37615475058555603,
-      "eval_mse": 0.1790776401758194,
-      "eval_runtime": 1.2719,
-      "eval_samples_per_second": 7.862,
-      "eval_steps_per_second": 1.572,
-      "step": 30
+      "eval_loss": 0.08366208523511887,
+      "eval_mae": 0.31394028663635254,
+      "eval_mse": 0.13212691247463226,
+      "eval_runtime": 9.4772,
+      "eval_samples_per_second": 8.019,
+      "eval_steps_per_second": 1.055,
+      "step": 150
     },
     {
       "epoch": 16.0,
-      "eval_loss": 0.08812335133552551,
-      "eval_mae": 0.38629305362701416,
-      "eval_mse": 0.1883399784564972,
-      "eval_runtime": 1.6525,
-      "eval_samples_per_second": 6.051,
-      "eval_steps_per_second": 1.21,
-      "step": 32
-    },
-    {
-      "epoch": 17.0,
-      "eval_loss": 0.09601505100727081,
-      "eval_mae": 0.38798943161964417,
-      "eval_mse": 0.18960143625736237,
-      "eval_runtime": 1.4237,
-      "eval_samples_per_second": 7.024,
-      "eval_steps_per_second": 1.405,
-      "step": 34
-    },
-    {
-      "epoch": 18.0,
-      "eval_loss": 0.0696040540933609,
-      "eval_mae": 0.3752928674221039,
-      "eval_mse": 0.17762857675552368,
-      "eval_runtime": 1.2854,
-      "eval_samples_per_second": 7.78,
-      "eval_steps_per_second": 1.556,
-      "step": 36
-    },
-    {
-      "epoch": 19.0,
-      "eval_loss": 0.057551633566617966,
-      "eval_mae": 0.3691748082637787,
-      "eval_mse": 0.1715679168701172,
-      "eval_runtime": 1.4944,
-      "eval_samples_per_second": 6.692,
-      "eval_steps_per_second": 1.338,
-      "step": 38
-    },
-    {
-      "epoch": 20.0,
-      "eval_loss": 0.05853904038667679,
-      "eval_mae": 0.3697223663330078,
-      "eval_mse": 0.17197829484939575,
-      "eval_runtime": 1.2655,
-      "eval_samples_per_second": 7.902,
-      "eval_steps_per_second": 1.58,
-      "step": 40
-    },
-    {
-      "epoch": 21.0,
-      "eval_loss": 0.07102437317371368,
-      "eval_mae": 0.37734654545783997,
-      "eval_mse": 0.1792123019695282,
-      "eval_runtime": 1.923,
-      "eval_samples_per_second": 5.2,
-      "eval_steps_per_second": 1.04,
-      "step": 42
-    },
-    {
-      "epoch": 22.0,
-      "eval_loss": 0.08145803213119507,
-      "eval_mae": 0.38286036252975464,
-      "eval_mse": 0.1842835247516632,
-      "eval_runtime": 1.1671,
-      "eval_samples_per_second": 8.568,
-      "eval_steps_per_second": 1.714,
-      "step": 44
-    },
-    {
-      "epoch": 23.0,
-      "eval_loss": 0.06858658790588379,
-      "eval_mae": 0.3709622919559479,
-      "eval_mse": 0.17373624444007874,
-      "eval_runtime": 1.3253,
-      "eval_samples_per_second": 7.545,
-      "eval_steps_per_second": 1.509,
-      "step": 46
-    },
-    {
-      "epoch": 24.0,
-      "eval_loss": 0.06739358603954315,
-      "eval_mae": 0.37041887640953064,
-      "eval_mse": 0.173396497964859,
-      "eval_runtime": 3.1433,
-      "eval_samples_per_second": 3.181,
-      "eval_steps_per_second": 0.636,
-      "step": 48
-    },
-    {
-      "epoch": 25.0,
-      "grad_norm": 0.9524662494659424,
-      "learning_rate": 1.53e-05,
-      "loss": 0.118,
-      "step": 50
-    },
-    {
-      "epoch": 25.0,
-      "eval_loss": 0.07070203125476837,
-      "eval_mae": 0.3756571412086487,
-      "eval_mse": 0.17759118974208832,
-      "eval_runtime": 1.2345,
-      "eval_samples_per_second": 8.1,
-      "eval_steps_per_second": 1.62,
-      "step": 50
-    },
-    {
-      "epoch": 26.0,
-      "eval_loss": 0.07530557364225388,
-      "eval_mae": 0.38040465116500854,
-      "eval_mse": 0.1816762387752533,
-      "eval_runtime": 1.7054,
-      "eval_samples_per_second": 5.864,
-      "eval_steps_per_second": 1.173,
-      "step": 52
-    },
-    {
-      "epoch": 27.0,
-      "eval_loss": 0.0708126351237297,
-      "eval_mae": 0.37711387872695923,
-      "eval_mse": 0.17874492704868317,
-      "eval_runtime": 1.1644,
-      "eval_samples_per_second": 8.588,
-      "eval_steps_per_second": 1.718,
-      "step": 54
+      "grad_norm": 2.5568466186523438,
+      "learning_rate": 1.1666666666666668e-05,
+      "loss": 0.2104,
+      "step": 160
     },
     {
-      "epoch": 28.0,
-      "eval_loss": 0.06371479481458664,
-      "eval_mae": 0.3699105381965637,
-      "eval_mse": 0.1721055954694748,
-      "eval_runtime": 1.5225,
-      "eval_samples_per_second": 6.568,
-      "eval_steps_per_second": 1.314,
-      "step": 56
-    },
-    {
-      "epoch": 29.0,
-      "eval_loss": 0.05992668867111206,
-      "eval_mae": 0.3671800494194031,
-      "eval_mse": 0.16954755783081055,
-      "eval_runtime": 1.3349,
-      "eval_samples_per_second": 7.491,
-      "eval_steps_per_second": 1.498,
-      "step": 58
-    },
-    {
-      "epoch": 30.0,
-      "eval_loss": 0.06269104778766632,
-      "eval_mae": 0.3710530698299408,
-      "eval_mse": 0.17309635877609253,
-      "eval_runtime": 1.2258,
-      "eval_samples_per_second": 8.158,
-      "eval_steps_per_second": 1.632,
-      "step": 60
-    },
-    {
-      "epoch": 31.0,
-      "eval_loss": 0.06929778307676315,
-      "eval_mae": 0.37701377272605896,
-      "eval_mse": 0.1785697340965271,
-      "eval_runtime": 1.4115,
-      "eval_samples_per_second": 7.085,
-      "eval_steps_per_second": 1.417,
-      "step": 62
-    },
-    {
-      "epoch": 32.0,
-      "eval_loss": 0.07379092276096344,
-      "eval_mae": 0.37811335921287537,
-      "eval_mse": 0.17990507185459137,
-      "eval_runtime": 1.2044,
-      "eval_samples_per_second": 8.303,
-      "eval_steps_per_second": 1.661,
-      "step": 64
-    },
-    {
-      "epoch": 33.0,
-      "eval_loss": 0.07304650545120239,
-      "eval_mae": 0.3751561939716339,
-      "eval_mse": 0.17731547355651855,
-      "eval_runtime": 3.3496,
-      "eval_samples_per_second": 2.985,
-      "eval_steps_per_second": 0.597,
-      "step": 66
-    },
-    {
-      "epoch": 34.0,
-      "eval_loss": 0.06837408244609833,
-      "eval_mae": 0.3710552453994751,
-      "eval_mse": 0.17353828251361847,
-      "eval_runtime": 1.4629,
-      "eval_samples_per_second": 6.836,
-      "eval_steps_per_second": 1.367,
-      "step": 68
-    },
-    {
-      "epoch": 35.0,
-      "eval_loss": 0.06416038423776627,
-      "eval_mae": 0.3673146069049835,
-      "eval_mse": 0.17018868029117584,
-      "eval_runtime": 1.31,
-      "eval_samples_per_second": 7.634,
-      "eval_steps_per_second": 1.527,
-      "step": 70
-    },
-    {
-      "epoch": 36.0,
-      "eval_loss": 0.06411489099264145,
-      "eval_mae": 0.36943870782852173,
-      "eval_mse": 0.1721121072769165,
-      "eval_runtime": 1.5048,
-      "eval_samples_per_second": 6.646,
-      "eval_steps_per_second": 1.329,
-      "step": 72
-    },
-    {
-      "epoch": 37.0,
-      "eval_loss": 0.06872588396072388,
-      "eval_mae": 0.373674213886261,
-      "eval_mse": 0.17579391598701477,
-      "eval_runtime": 1.4802,
-      "eval_samples_per_second": 6.756,
-      "eval_steps_per_second": 1.351,
-      "step": 74
-    },
-    {
-      "epoch": 38.0,
-      "eval_loss": 0.07391238957643509,
-      "eval_mae": 0.3772023916244507,
-      "eval_mse": 0.17880253493785858,
-      "eval_runtime": 1.3549,
-      "eval_samples_per_second": 7.381,
-      "eval_steps_per_second": 1.476,
-      "step": 76
-    },
-    {
-      "epoch": 39.0,
-      "eval_loss": 0.07059959322214127,
-      "eval_mae": 0.3748144209384918,
-      "eval_mse": 0.1765395551919937,
-      "eval_runtime": 1.3015,
-      "eval_samples_per_second": 7.683,
-      "eval_steps_per_second": 1.537,
-      "step": 78
-    },
-    {
-      "epoch": 40.0,
-      "eval_loss": 0.06648191064596176,
-      "eval_mae": 0.37035736441612244,
-      "eval_mse": 0.17257149517536163,
-      "eval_runtime": 1.4739,
-      "eval_samples_per_second": 6.785,
-      "eval_steps_per_second": 1.357,
-      "step": 80
-    },
-    {
-      "epoch": 41.0,
-      "eval_loss": 0.06419553607702255,
-      "eval_mae": 0.36771345138549805,
-      "eval_mse": 0.1702636033296585,
-      "eval_runtime": 1.3959,
-      "eval_samples_per_second": 7.164,
-      "eval_steps_per_second": 1.433,
-      "step": 82
-    },
-    {
-      "epoch": 42.0,
-      "eval_loss": 0.06590329110622406,
-      "eval_mae": 0.3694536089897156,
-      "eval_mse": 0.17185601592063904,
-      "eval_runtime": 1.3544,
-      "eval_samples_per_second": 7.383,
-      "eval_steps_per_second": 1.477,
-      "step": 84
-    },
-    {
-      "epoch": 43.0,
-      "eval_loss": 0.06824100762605667,
-      "eval_mae": 0.3721493184566498,
-      "eval_mse": 0.1743040680885315,
-      "eval_runtime": 1.3016,
-      "eval_samples_per_second": 7.683,
-      "eval_steps_per_second": 1.537,
-      "step": 86
-    },
-    {
-      "epoch": 44.0,
-      "eval_loss": 0.07233113795518875,
-      "eval_mae": 0.37587830424308777,
-      "eval_mse": 0.17762817442417145,
-      "eval_runtime": 1.2282,
-      "eval_samples_per_second": 8.142,
-      "eval_steps_per_second": 1.628,
-      "step": 88
-    },
-    {
-      "epoch": 45.0,
-      "eval_loss": 0.0730481892824173,
-      "eval_mae": 0.37610477209091187,
-      "eval_mse": 0.1778472512960434,
-      "eval_runtime": 1.2255,
-      "eval_samples_per_second": 8.16,
-      "eval_steps_per_second": 1.632,
-      "step": 90
-    },
-    {
-      "epoch": 46.0,
-      "eval_loss": 0.07376444339752197,
-      "eval_mae": 0.37569037079811096,
-      "eval_mse": 0.17753194272518158,
-      "eval_runtime": 1.2758,
-      "eval_samples_per_second": 7.838,
-      "eval_steps_per_second": 1.568,
-      "step": 92
-    },
-    {
-      "epoch": 47.0,
-      "eval_loss": 0.07429193705320358,
-      "eval_mae": 0.3751032054424286,
-      "eval_mse": 0.1769881248474121,
-      "eval_runtime": 1.1953,
-      "eval_samples_per_second": 8.366,
-      "eval_steps_per_second": 1.673,
-      "step": 94
-    },
-    {
-      "epoch": 48.0,
-      "eval_loss": 0.07430405914783478,
-      "eval_mae": 0.3747301399707794,
-      "eval_mse": 0.17660388350486755,
-      "eval_runtime": 1.1671,
-      "eval_samples_per_second": 8.568,
-      "eval_steps_per_second": 1.714,
-      "step": 96
-    },
-    {
-      "epoch": 49.0,
-      "eval_loss": 0.07368143647909164,
-      "eval_mae": 0.37456196546554565,
-      "eval_mse": 0.17642483115196228,
-      "eval_runtime": 1.545,
-      "eval_samples_per_second": 6.472,
-      "eval_steps_per_second": 1.294,
-      "step": 98
+      "epoch": 16.0,
+      "eval_loss": 0.09077440947294235,
+      "eval_mae": 0.3225802481174469,
+      "eval_mse": 0.14043577015399933,
+      "eval_runtime": 9.6623,
+      "eval_samples_per_second": 7.866,
+      "eval_steps_per_second": 1.035,
+      "step": 160
     },
     {
-      "epoch": 50.0,
-      "grad_norm": 0.7434751987457275,
-      "learning_rate": 3.0000000000000004e-07,
-      "loss": 0.0449,
-      "step": 100
+      "epoch": 17.0,
+      "eval_loss": 0.08574231714010239,
+      "eval_mae": 0.31825631856918335,
+      "eval_mse": 0.13667234778404236,
+      "eval_runtime": 9.4595,
+      "eval_samples_per_second": 8.034,
+      "eval_steps_per_second": 1.057,
+      "step": 170
     },
     {
-      "epoch": 50.0,
-      "eval_loss": 0.07292107492685318,
-      "eval_mae": 0.3741002678871155,
-      "eval_mse": 0.17599670588970184,
-      "eval_runtime": 1.2607,
-      "eval_samples_per_second": 7.932,
-      "eval_steps_per_second": 1.586,
-      "step": 100
+      "epoch": 18.0,
+      "eval_loss": 0.08340632170438766,
+      "eval_mae": 0.3171162009239197,
+      "eval_mse": 0.13509555160999298,
+      "eval_runtime": 9.4982,
+      "eval_samples_per_second": 8.002,
+      "eval_steps_per_second": 1.053,
+      "step": 180
     },
     {
-      "epoch": 50.0,
-      "step": 100,
+      "epoch": 18.0,
+      "step": 180,
       "total_flos": 0.0,
-      "train_loss": 0.08142914772033691,
-      "train_runtime": 429.9896,
-      "train_samples_per_second": 9.07,
-      "train_steps_per_second": 0.233
+      "train_loss": 0.2600840449333191,
+      "train_runtime": 1002.0455,
+      "train_samples_per_second": 6.048,
+      "train_steps_per_second": 0.18
     }
   ],
-  "logging_steps": 50,
-  "max_steps": 100,
+  "logging_steps": 40,
+  "max_steps": 180,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 100,
+  "num_train_epochs": 20,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {