End of training

README.md

@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [apple/mobilevit-xx-small](https://huggingface.co/apple/mobilevit-xx-small) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0688
-- Accuracy: 0.64
+- Loss: 1.0321
+- Accuracy: 0.66
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 30.0,
-    "eval_accuracy": 0.8922222222222222,
-    "eval_loss": 0.5839348435401917,
-    "eval_runtime": 641.5157,
-    "eval_samples_per_second": 4.209,
-    "eval_steps_per_second": 0.132,
-    "total_flos": 2.136995592192e+16,
-    "train_loss": 1.1791712323824564,
-    "train_runtime": 995.7979,
-    "train_samples_per_second": 15.063,
-    "train_steps_per_second": 0.121
+    "epoch": 46.15384615384615,
+    "eval_accuracy": 0.66,
+    "eval_loss": 1.0320532321929932,
+    "eval_runtime": 0.4192,
+    "eval_samples_per_second": 238.552,
+    "eval_steps_per_second": 9.542,
+    "total_flos": 2.630499107615539e+16,
+    "train_loss": 1.076003630956014,
+    "train_runtime": 895.7496,
+    "train_samples_per_second": 22.328,
+    "train_steps_per_second": 0.167
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.8922222222222222,
-    "eval_loss": 0.5839348435401917,
-    "eval_runtime": 641.5157,
-    "eval_samples_per_second": 4.209,
-    "eval_steps_per_second": 0.132
+    "epoch": 46.15384615384615,
+    "eval_accuracy": 0.66,
+    "eval_loss": 1.0320532321929932,
+    "eval_runtime": 0.4192,
+    "eval_samples_per_second": 238.552,
+    "eval_steps_per_second": 9.542
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 30.0,
-    "total_flos": 2.136995592192e+16,
-    "train_loss": 1.1791712323824564,
-    "train_runtime": 995.7979,
-    "train_samples_per_second": 15.063,
-    "train_steps_per_second": 0.121
+    "epoch": 46.15384615384615,
+    "total_flos": 2.630499107615539e+16,
+    "train_loss": 1.076003630956014,
+    "train_runtime": 895.7496,
+    "train_samples_per_second": 22.328,
+    "train_steps_per_second": 0.167
 }

trainer_state.json

@@ -1,381 +1,555 @@
 {
-  "best_metric": 0.636,
-  "best_model_checkpoint": "mobilevit-xx-small-finetuned-eurosat/checkpoint-120",
-  "epoch": 30.0,
+  "best_metric": 0.66,
+  "best_model_checkpoint": "mobilevit-xx-small-finetuned-eurosat/checkpoint-94",
+  "epoch": 46.15384615384615,
   "eval_steps": 500,
-  "global_step": 120,
+  "global_step": 150,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "eval_accuracy": 0.212,
-      "eval_loss": 2.2717537879943848,
-      "eval_runtime": 11.4285,
-      "eval_samples_per_second": 43.75,
-      "eval_steps_per_second": 1.4,
-      "step": 4
+      "epoch": 0.9230769230769231,
+      "eval_accuracy": 0.11,
+      "eval_loss": 2.2814395427703857,
+      "eval_runtime": 0.3447,
+      "eval_samples_per_second": 290.129,
+      "eval_steps_per_second": 11.605,
+      "step": 3
     },
     {
-      "epoch": 2.0,
-      "eval_accuracy": 0.29,
-      "eval_loss": 2.1031904220581055,
-      "eval_runtime": 20.4802,
-      "eval_samples_per_second": 24.414,
-      "eval_steps_per_second": 0.781,
-      "step": 8
+      "epoch": 1.8461538461538463,
+      "eval_accuracy": 0.26,
+      "eval_loss": 2.227895736694336,
+      "eval_runtime": 0.4326,
+      "eval_samples_per_second": 231.174,
+      "eval_steps_per_second": 9.247,
+      "step": 6
     },
     {
-      "epoch": 2.5,
-      "grad_norm": 4.617868423461914,
-      "learning_rate": 0.0025,
-      "loss": 2.1883,
-      "step": 10
+      "epoch": 2.769230769230769,
+      "eval_accuracy": 0.29,
+      "eval_loss": 2.105987071990967,
+      "eval_runtime": 0.4383,
+      "eval_samples_per_second": 228.177,
+      "eval_steps_per_second": 9.127,
+      "step": 9
     },
     {
-      "epoch": 3.0,
-      "eval_accuracy": 0.336,
-      "eval_loss": 1.9200304746627808,
-      "eval_runtime": 2.4299,
-      "eval_samples_per_second": 205.771,
-      "eval_steps_per_second": 6.585,
-      "step": 12
+      "epoch": 3.076923076923077,
+      "grad_norm": 3.6058764457702637,
+      "learning_rate": 0.002,
+      "loss": 2.2256,
+      "step": 10
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.26,
-      "eval_loss": 2.733245849609375,
-      "eval_runtime": 1.9139,
-      "eval_samples_per_second": 261.247,
-      "eval_steps_per_second": 8.36,
+      "eval_accuracy": 0.31,
+      "eval_loss": 1.8600536584854126,
+      "eval_runtime": 12.3694,
+      "eval_samples_per_second": 8.084,
+      "eval_steps_per_second": 0.323,
+      "step": 13
+    },
+    {
+      "epoch": 4.923076923076923,
+      "eval_accuracy": 0.32,
+      "eval_loss": 1.8917309045791626,
+      "eval_runtime": 13.8701,
+      "eval_samples_per_second": 7.21,
+      "eval_steps_per_second": 0.288,
       "step": 16
     },
     {
-      "epoch": 5.0,
-      "grad_norm": 7.1269941329956055,
-      "learning_rate": 0.002777777777777778,
-      "loss": 1.7537,
-      "step": 20
+      "epoch": 5.846153846153846,
+      "eval_accuracy": 0.33,
+      "eval_loss": 1.7589596509933472,
+      "eval_runtime": 14.8817,
+      "eval_samples_per_second": 6.72,
+      "eval_steps_per_second": 0.269,
+      "step": 19
     },
     {
-      "epoch": 5.0,
-      "eval_accuracy": 0.354,
-      "eval_loss": 2.0565059185028076,
-      "eval_runtime": 22.3861,
-      "eval_samples_per_second": 22.335,
-      "eval_steps_per_second": 0.715,
+      "epoch": 6.153846153846154,
+      "grad_norm": 3.9451377391815186,
+      "learning_rate": 0.0028888888888888888,
+      "loss": 1.8,
       "step": 20
     },
     {
-      "epoch": 6.0,
-      "eval_accuracy": 0.352,
-      "eval_loss": 1.9995957612991333,
-      "eval_runtime": 1.94,
-      "eval_samples_per_second": 257.729,
-      "eval_steps_per_second": 8.247,
-      "step": 24
-    },
-    {
-      "epoch": 7.0,
-      "eval_accuracy": 0.388,
-      "eval_loss": 1.7918899059295654,
-      "eval_runtime": 21.978,
-      "eval_samples_per_second": 22.75,
-      "eval_steps_per_second": 0.728,
-      "step": 28
+      "epoch": 6.769230769230769,
+      "eval_accuracy": 0.21,
+      "eval_loss": 2.9505045413970947,
+      "eval_runtime": 0.8737,
+      "eval_samples_per_second": 114.459,
+      "eval_steps_per_second": 4.578,
+      "step": 22
     },
     {
-      "epoch": 7.5,
-      "grad_norm": 3.362823247909546,
-      "learning_rate": 0.0025,
-      "loss": 1.525,
+      "epoch": 8.0,
+      "eval_accuracy": 0.19,
+      "eval_loss": 2.440716505050659,
+      "eval_runtime": 0.3849,
+      "eval_samples_per_second": 259.81,
+      "eval_steps_per_second": 10.392,
+      "step": 26
+    },
+    {
+      "epoch": 8.923076923076923,
+      "eval_accuracy": 0.36,
+      "eval_loss": 2.6913070678710938,
+      "eval_runtime": 0.4553,
+      "eval_samples_per_second": 219.638,
+      "eval_steps_per_second": 8.786,
+      "step": 29
+    },
+    {
+      "epoch": 9.23076923076923,
+      "grad_norm": 5.145291805267334,
+      "learning_rate": 0.0026666666666666666,
+      "loss": 1.6404,
       "step": 30
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.366,
-      "eval_loss": 1.9682480096817017,
-      "eval_runtime": 22.1475,
-      "eval_samples_per_second": 22.576,
-      "eval_steps_per_second": 0.722,
+      "epoch": 9.846153846153847,
+      "eval_accuracy": 0.48,
+      "eval_loss": 1.409439206123352,
+      "eval_runtime": 16.0101,
+      "eval_samples_per_second": 6.246,
+      "eval_steps_per_second": 0.25,
       "step": 32
     },
     {
-      "epoch": 9.0,
-      "eval_accuracy": 0.44,
-      "eval_loss": 1.550385594367981,
-      "eval_runtime": 4.0809,
-      "eval_samples_per_second": 122.521,
-      "eval_steps_per_second": 3.921,
-      "step": 36
+      "epoch": 10.76923076923077,
+      "eval_accuracy": 0.47,
+      "eval_loss": 1.401109218597412,
+      "eval_runtime": 1.2056,
+      "eval_samples_per_second": 82.944,
+      "eval_steps_per_second": 3.318,
+      "step": 35
     },
     {
-      "epoch": 10.0,
-      "grad_norm": 6.439978122711182,
-      "learning_rate": 0.0022222222222222222,
-      "loss": 1.41,
+      "epoch": 12.0,
+      "eval_accuracy": 0.45,
+      "eval_loss": 1.759374737739563,
+      "eval_runtime": 16.745,
+      "eval_samples_per_second": 5.972,
+      "eval_steps_per_second": 0.239,
+      "step": 39
+    },
+    {
+      "epoch": 12.307692307692308,
+      "grad_norm": 9.834226608276367,
+      "learning_rate": 0.0024444444444444444,
+      "loss": 1.4817,
       "step": 40
     },
     {
-      "epoch": 10.0,
-      "eval_accuracy": 0.406,
-      "eval_loss": 1.6390491724014282,
-      "eval_runtime": 15.0604,
-      "eval_samples_per_second": 33.2,
-      "eval_steps_per_second": 1.062,
-      "step": 40
+      "epoch": 12.923076923076923,
+      "eval_accuracy": 0.46,
+      "eval_loss": 1.431288719177246,
+      "eval_runtime": 0.4722,
+      "eval_samples_per_second": 211.785,
+      "eval_steps_per_second": 8.471,
+      "step": 42
     },
     {
-      "epoch": 11.0,
-      "eval_accuracy": 0.482,
-      "eval_loss": 1.8054633140563965,
-      "eval_runtime": 22.4978,
-      "eval_samples_per_second": 22.224,
-      "eval_steps_per_second": 0.711,
-      "step": 44
+      "epoch": 13.846153846153847,
+      "eval_accuracy": 0.49,
+      "eval_loss": 1.2744395732879639,
+      "eval_runtime": 6.1902,
+      "eval_samples_per_second": 16.155,
+      "eval_steps_per_second": 0.646,
+      "step": 45
     },
     {
-      "epoch": 12.0,
-      "eval_accuracy": 0.372,
-      "eval_loss": 1.8536120653152466,
-      "eval_runtime": 1.8645,
-      "eval_samples_per_second": 268.166,
-      "eval_steps_per_second": 8.581,
+      "epoch": 14.76923076923077,
+      "eval_accuracy": 0.46,
+      "eval_loss": 1.4527015686035156,
+      "eval_runtime": 0.337,
+      "eval_samples_per_second": 296.7,
+      "eval_steps_per_second": 11.868,
       "step": 48
     },
     {
-      "epoch": 12.5,
-      "grad_norm": 4.127369403839111,
-      "learning_rate": 0.0019444444444444444,
-      "loss": 1.2638,
+      "epoch": 15.384615384615385,
+      "grad_norm": 3.165754795074463,
+      "learning_rate": 0.0022222222222222222,
+      "loss": 1.2872,
       "step": 50
     },
     {
-      "epoch": 13.0,
-      "eval_accuracy": 0.488,
-      "eval_loss": 1.6079543828964233,
-      "eval_runtime": 1.8468,
-      "eval_samples_per_second": 270.739,
-      "eval_steps_per_second": 8.664,
+      "epoch": 16.0,
+      "eval_accuracy": 0.42,
+      "eval_loss": 1.5522328615188599,
+      "eval_runtime": 0.4566,
+      "eval_samples_per_second": 219.028,
+      "eval_steps_per_second": 8.761,
       "step": 52
     },
     {
-      "epoch": 14.0,
-      "eval_accuracy": 0.418,
-      "eval_loss": 1.810085415840149,
-      "eval_runtime": 21.9085,
-      "eval_samples_per_second": 22.822,
-      "eval_steps_per_second": 0.73,
-      "step": 56
+      "epoch": 16.923076923076923,
+      "eval_accuracy": 0.38,
+      "eval_loss": 1.7599867582321167,
+      "eval_runtime": 15.1941,
+      "eval_samples_per_second": 6.582,
+      "eval_steps_per_second": 0.263,
+      "step": 55
     },
     {
-      "epoch": 15.0,
-      "grad_norm": 5.039086818695068,
-      "learning_rate": 0.0016666666666666668,
-      "loss": 1.1194,
-      "step": 60
+      "epoch": 17.846153846153847,
+      "eval_accuracy": 0.53,
+      "eval_loss": 1.3952364921569824,
+      "eval_runtime": 0.4698,
+      "eval_samples_per_second": 212.838,
+      "eval_steps_per_second": 8.514,
+      "step": 58
     },
     {
-      "epoch": 15.0,
-      "eval_accuracy": 0.422,
-      "eval_loss": 1.742574691772461,
-      "eval_runtime": 22.139,
-      "eval_samples_per_second": 22.585,
-      "eval_steps_per_second": 0.723,
+      "epoch": 18.46153846153846,
+      "grad_norm": 3.935255765914917,
+      "learning_rate": 0.002,
+      "loss": 1.2307,
       "step": 60
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.452,
-      "eval_loss": 1.7981486320495605,
-      "eval_runtime": 39.1443,
-      "eval_samples_per_second": 12.773,
-      "eval_steps_per_second": 0.409,
-      "step": 64
-    },
-    {
-      "epoch": 17.0,
-      "eval_accuracy": 0.392,
-      "eval_loss": 2.088580369949341,
-      "eval_runtime": 1.8333,
-      "eval_samples_per_second": 272.732,
-      "eval_steps_per_second": 8.727,
+      "epoch": 18.76923076923077,
+      "eval_accuracy": 0.51,
+      "eval_loss": 1.5067017078399658,
+      "eval_runtime": 8.3399,
+      "eval_samples_per_second": 11.991,
+      "eval_steps_per_second": 0.48,
+      "step": 61
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": 0.51,
+      "eval_loss": 1.631130576133728,
+      "eval_runtime": 14.6358,
+      "eval_samples_per_second": 6.833,
+      "eval_steps_per_second": 0.273,
+      "step": 65
+    },
+    {
+      "epoch": 20.923076923076923,
+      "eval_accuracy": 0.58,
+      "eval_loss": 1.3986690044403076,
+      "eval_runtime": 0.4557,
+      "eval_samples_per_second": 219.438,
+      "eval_steps_per_second": 8.778,
       "step": 68
     },
     {
-      "epoch": 17.5,
-      "grad_norm": 3.298395872116089,
-      "learning_rate": 0.001388888888888889,
-      "loss": 1.0082,
+      "epoch": 21.53846153846154,
+      "grad_norm": 6.055632591247559,
+      "learning_rate": 0.0017777777777777776,
+      "loss": 1.0604,
       "step": 70
     },
     {
-      "epoch": 18.0,
-      "eval_accuracy": 0.538,
-      "eval_loss": 1.4252102375030518,
-      "eval_runtime": 2.3763,
-      "eval_samples_per_second": 210.41,
-      "eval_steps_per_second": 6.733,
-      "step": 72
+      "epoch": 21.846153846153847,
+      "eval_accuracy": 0.54,
+      "eval_loss": 1.4127693176269531,
+      "eval_runtime": 0.4206,
+      "eval_samples_per_second": 237.781,
+      "eval_steps_per_second": 9.511,
+      "step": 71
     },
     {
-      "epoch": 19.0,
-      "eval_accuracy": 0.576,
-      "eval_loss": 1.3337562084197998,
-      "eval_runtime": 2.0774,
-      "eval_samples_per_second": 240.688,
-      "eval_steps_per_second": 7.702,
-      "step": 76
+      "epoch": 22.76923076923077,
+      "eval_accuracy": 0.56,
+      "eval_loss": 1.1939380168914795,
+      "eval_runtime": 0.4239,
+      "eval_samples_per_second": 235.924,
+      "eval_steps_per_second": 9.437,
+      "step": 74
     },
     {
-      "epoch": 20.0,
-      "grad_norm": 3.3816585540771484,
-      "learning_rate": 0.0011111111111111111,
-      "loss": 0.9556,
+      "epoch": 24.0,
+      "eval_accuracy": 0.49,
+      "eval_loss": 1.4323681592941284,
+      "eval_runtime": 0.3707,
+      "eval_samples_per_second": 269.772,
+      "eval_steps_per_second": 10.791,
+      "step": 78
+    },
+    {
+      "epoch": 24.615384615384617,
+      "grad_norm": 8.553020477294922,
+      "learning_rate": 0.0015555555555555555,
+      "loss": 1.0053,
       "step": 80
     },
     {
-      "epoch": 20.0,
-      "eval_accuracy": 0.492,
-      "eval_loss": 1.717270851135254,
-      "eval_runtime": 1.9665,
-      "eval_samples_per_second": 254.261,
-      "eval_steps_per_second": 8.136,
-      "step": 80
+      "epoch": 24.923076923076923,
+      "eval_accuracy": 0.53,
+      "eval_loss": 1.3661165237426758,
+      "eval_runtime": 0.4573,
+      "eval_samples_per_second": 218.668,
+      "eval_steps_per_second": 8.747,
+      "step": 81
     },
     {
-      "epoch": 21.0,
-      "eval_accuracy": 0.572,
-      "eval_loss": 1.4067368507385254,
-      "eval_runtime": 22.3773,
-      "eval_samples_per_second": 22.344,
-      "eval_steps_per_second": 0.715,
+      "epoch": 25.846153846153847,
+      "eval_accuracy": 0.53,
+      "eval_loss": 1.25278902053833,
+      "eval_runtime": 0.3451,
+      "eval_samples_per_second": 289.798,
+      "eval_steps_per_second": 11.592,
       "step": 84
     },
     {
-      "epoch": 22.0,
-      "eval_accuracy": 0.614,
-      "eval_loss": 1.1839243173599243,
-      "eval_runtime": 1.8497,
-      "eval_samples_per_second": 270.312,
-      "eval_steps_per_second": 8.65,
-      "step": 88
+      "epoch": 26.76923076923077,
+      "eval_accuracy": 0.57,
+      "eval_loss": 1.204034447669983,
+      "eval_runtime": 2.5122,
+      "eval_samples_per_second": 39.806,
+      "eval_steps_per_second": 1.592,
+      "step": 87
     },
     {
-      "epoch": 22.5,
-      "grad_norm": 4.263976097106934,
-      "learning_rate": 0.0008333333333333334,
-      "loss": 0.8404,
+      "epoch": 27.692307692307693,
+      "grad_norm": 3.5659632682800293,
+      "learning_rate": 0.0013333333333333333,
+      "loss": 0.8327,
       "step": 90
     },
     {
-      "epoch": 23.0,
-      "eval_accuracy": 0.612,
-      "eval_loss": 1.2036840915679932,
-      "eval_runtime": 1.8642,
-      "eval_samples_per_second": 268.207,
-      "eval_steps_per_second": 8.583,
-      "step": 92
+      "epoch": 28.0,
+      "eval_accuracy": 0.61,
+      "eval_loss": 1.1885842084884644,
+      "eval_runtime": 0.5132,
+      "eval_samples_per_second": 194.851,
+      "eval_steps_per_second": 7.794,
+      "step": 91
     },
     {
-      "epoch": 24.0,
-      "eval_accuracy": 0.552,
-      "eval_loss": 1.4832416772842407,
-      "eval_runtime": 1.9785,
-      "eval_samples_per_second": 252.716,
-      "eval_steps_per_second": 8.087,
-      "step": 96
-    },
-    {
-      "epoch": 25.0,
-      "grad_norm": 3.777881145477295,
-      "learning_rate": 0.0005555555555555556,
-      "loss": 0.7603,
-      "step": 100
+      "epoch": 28.923076923076923,
+      "eval_accuracy": 0.66,
+      "eval_loss": 1.0320532321929932,
+      "eval_runtime": 17.1732,
+      "eval_samples_per_second": 5.823,
+      "eval_steps_per_second": 0.233,
+      "step": 94
+    },
+    {
+      "epoch": 29.846153846153847,
+      "eval_accuracy": 0.61,
+      "eval_loss": 1.1496257781982422,
+      "eval_runtime": 0.4295,
+      "eval_samples_per_second": 232.828,
+      "eval_steps_per_second": 9.313,
+      "step": 97
     },
     {
-      "epoch": 25.0,
-      "eval_accuracy": 0.592,
-      "eval_loss": 1.3205021619796753,
-      "eval_runtime": 1.9133,
-      "eval_samples_per_second": 261.322,
-      "eval_steps_per_second": 8.362,
+      "epoch": 30.76923076923077,
+      "grad_norm": 6.044861793518066,
+      "learning_rate": 0.0011111111111111111,
+      "loss": 0.7456,
       "step": 100
     },
     {
-      "epoch": 26.0,
+      "epoch": 30.76923076923077,
       "eval_accuracy": 0.61,
-      "eval_loss": 1.250473976135254,
-      "eval_runtime": 41.202,
-      "eval_samples_per_second": 12.135,
-      "eval_steps_per_second": 0.388,
+      "eval_loss": 1.1801018714904785,
+      "eval_runtime": 0.8649,
+      "eval_samples_per_second": 115.616,
+      "eval_steps_per_second": 4.625,
+      "step": 100
+    },
+    {
+      "epoch": 32.0,
+      "eval_accuracy": 0.53,
+      "eval_loss": 1.288604736328125,
+      "eval_runtime": 0.3454,
+      "eval_samples_per_second": 289.55,
+      "eval_steps_per_second": 11.582,
       "step": 104
     },
     {
-      "epoch": 27.0,
-      "eval_accuracy": 0.622,
-      "eval_loss": 1.2412116527557373,
-      "eval_runtime": 1.9569,
-      "eval_samples_per_second": 255.501,
-      "eval_steps_per_second": 8.176,
-      "step": 108
+      "epoch": 32.92307692307692,
+      "eval_accuracy": 0.6,
+      "eval_loss": 1.2215436697006226,
+      "eval_runtime": 0.3906,
+      "eval_samples_per_second": 256.03,
+      "eval_steps_per_second": 10.241,
+      "step": 107
     },
     {
-      "epoch": 27.5,
-      "grad_norm": 3.8510005474090576,
-      "learning_rate": 0.0002777777777777778,
-      "loss": 0.7096,
+      "epoch": 33.84615384615385,
+      "grad_norm": 6.93057107925415,
+      "learning_rate": 0.0008888888888888888,
+      "loss": 0.7106,
       "step": 110
     },
     {
-      "epoch": 28.0,
-      "eval_accuracy": 0.626,
-      "eval_loss": 1.1740450859069824,
-      "eval_runtime": 6.5344,
-      "eval_samples_per_second": 76.518,
-      "eval_steps_per_second": 2.449,
-      "step": 112
+      "epoch": 33.84615384615385,
+      "eval_accuracy": 0.55,
+      "eval_loss": 1.2372339963912964,
+      "eval_runtime": 11.0141,
+      "eval_samples_per_second": 9.079,
+      "eval_steps_per_second": 0.363,
+      "step": 110
     },
     {
-      "epoch": 29.0,
-      "eval_accuracy": 0.624,
-      "eval_loss": 1.1464664936065674,
-      "eval_runtime": 20.8737,
-      "eval_samples_per_second": 23.954,
-      "eval_steps_per_second": 0.767,
-      "step": 116
+      "epoch": 34.76923076923077,
+      "eval_accuracy": 0.62,
+      "eval_loss": 1.1833901405334473,
+      "eval_runtime": 0.3401,
+      "eval_samples_per_second": 294.013,
+      "eval_steps_per_second": 11.761,
+      "step": 113
     },
     {
-      "epoch": 30.0,
-      "grad_norm": 3.827312707901001,
-      "learning_rate": 0.0,
-      "loss": 0.6157,
+      "epoch": 36.0,
+      "eval_accuracy": 0.64,
+      "eval_loss": 1.2001395225524902,
+      "eval_runtime": 0.4622,
+      "eval_samples_per_second": 216.349,
+      "eval_steps_per_second": 8.654,
+      "step": 117
+    },
+    {
+      "epoch": 36.92307692307692,
+      "grad_norm": 4.3674445152282715,
+      "learning_rate": 0.0006666666666666666,
+      "loss": 0.5914,
       "step": 120
     },
     {
-      "epoch": 30.0,
-      "eval_accuracy": 0.636,
-      "eval_loss": 1.1490942239761353,
-      "eval_runtime": 17.8767,
-      "eval_samples_per_second": 27.969,
-      "eval_steps_per_second": 0.895,
+      "epoch": 36.92307692307692,
+      "eval_accuracy": 0.63,
+      "eval_loss": 1.1181855201721191,
+      "eval_runtime": 8.8978,
+      "eval_samples_per_second": 11.239,
+      "eval_steps_per_second": 0.45,
       "step": 120
     },
     {
-      "epoch": 30.0,
-      "step": 120,
-      "total_flos": 2.136995592192e+16,
-      "train_loss": 1.1791712323824564,
-      "train_runtime": 995.7979,
-      "train_samples_per_second": 15.063,
-      "train_steps_per_second": 0.121
+      "epoch": 37.84615384615385,
+      "eval_accuracy": 0.62,
+      "eval_loss": 1.2223446369171143,
+      "eval_runtime": 0.4156,
+      "eval_samples_per_second": 240.596,
+      "eval_steps_per_second": 9.624,
+      "step": 123
+    },
+    {
+      "epoch": 38.76923076923077,
+      "eval_accuracy": 0.59,
+      "eval_loss": 1.244262456893921,
+      "eval_runtime": 0.4328,
+      "eval_samples_per_second": 231.05,
+      "eval_steps_per_second": 9.242,
+      "step": 126
+    },
+    {
+      "epoch": 40.0,
+      "grad_norm": 3.360128402709961,
+      "learning_rate": 0.0004444444444444444,
+      "loss": 0.5601,
+      "step": 130
+    },
+    {
+      "epoch": 40.0,
+      "eval_accuracy": 0.59,
+      "eval_loss": 1.2605319023132324,
+      "eval_runtime": 13.3957,
+      "eval_samples_per_second": 7.465,
+      "eval_steps_per_second": 0.299,
+      "step": 130
+    },
+    {
+      "epoch": 40.92307692307692,
+      "eval_accuracy": 0.6,
+      "eval_loss": 1.2939895391464233,
+      "eval_runtime": 5.4663,
+      "eval_samples_per_second": 18.294,
+      "eval_steps_per_second": 0.732,
+      "step": 133
+    },
+    {
+      "epoch": 41.84615384615385,
+      "eval_accuracy": 0.59,
+      "eval_loss": 1.2283549308776855,
+      "eval_runtime": 0.4128,
+      "eval_samples_per_second": 242.228,
+      "eval_steps_per_second": 9.689,
+      "step": 136
+    },
+    {
+      "epoch": 42.76923076923077,
+      "eval_accuracy": 0.61,
+      "eval_loss": 1.1883938312530518,
+      "eval_runtime": 0.3379,
+      "eval_samples_per_second": 295.934,
+      "eval_steps_per_second": 11.837,
+      "step": 139
+    },
+    {
+      "epoch": 43.07692307692308,
+      "grad_norm": 7.6419358253479,
+      "learning_rate": 0.0002222222222222222,
+      "loss": 0.5151,
+      "step": 140
+    },
+    {
+      "epoch": 44.0,
+      "eval_accuracy": 0.63,
+      "eval_loss": 1.1789089441299438,
+      "eval_runtime": 0.4065,
+      "eval_samples_per_second": 245.995,
+      "eval_steps_per_second": 9.84,
+      "step": 143
+    },
+    {
+      "epoch": 44.92307692307692,
+      "eval_accuracy": 0.62,
+      "eval_loss": 1.129704475402832,
+      "eval_runtime": 0.4164,
+      "eval_samples_per_second": 240.144,
+      "eval_steps_per_second": 9.606,
+      "step": 146
+    },
+    {
+      "epoch": 45.84615384615385,
+      "eval_accuracy": 0.63,
+      "eval_loss": 1.0879005193710327,
+      "eval_runtime": 0.6008,
+      "eval_samples_per_second": 166.444,
+      "eval_steps_per_second": 6.658,
+      "step": 149
+    },
+    {
+      "epoch": 46.15384615384615,
+      "grad_norm": 2.6147122383117676,
+      "learning_rate": 0.0,
+      "loss": 0.4531,
+      "step": 150
+    },
+    {
+      "epoch": 46.15384615384615,
+      "eval_accuracy": 0.64,
+      "eval_loss": 1.068832516670227,
+      "eval_runtime": 8.4147,
+      "eval_samples_per_second": 11.884,
+      "eval_steps_per_second": 0.475,
+      "step": 150
+    },
+    {
+      "epoch": 46.15384615384615,
+      "step": 150,
+      "total_flos": 2.630499107615539e+16,
+      "train_loss": 1.076003630956014,
+      "train_runtime": 895.7496,
+      "train_samples_per_second": 22.328,
+      "train_steps_per_second": 0.167
     }
   ],
   "logging_steps": 10,
-  "max_steps": 120,
+  "max_steps": 150,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 30,
+  "num_train_epochs": 50,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -389,7 +563,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.136995592192e+16,
+  "total_flos": 2.630499107615539e+16,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null