End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 4.98,
-    "eval_accuracy": 0.9775132275132276,
-    "eval_loss": 0.09184322506189346,
-    "eval_runtime": 4.8893,
-    "eval_samples_per_second": 154.623,
-    "eval_steps_per_second": 4.909,
-    "total_flos": 5.346226696020296e+17,
-    "train_loss": 0.37407838611891775,
-    "train_runtime": 386.1594,
-    "train_samples_per_second": 55.935,
-    "train_steps_per_second": 0.427
+    "epoch": 9.98,
+    "eval_accuracy": 0.9828042328042328,
+    "eval_loss": 0.06841638684272766,
+    "eval_runtime": 5.801,
+    "eval_samples_per_second": 130.321,
+    "eval_steps_per_second": 4.137,
+    "total_flos": 1.0716320475504968e+18,
+    "train_loss": 0.24182906033414783,
+    "train_runtime": 827.7733,
+    "train_samples_per_second": 52.188,
+    "train_steps_per_second": 0.399
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.98,
-    "eval_accuracy": 0.9775132275132276,
-    "eval_loss": 0.09184322506189346,
-    "eval_runtime": 4.8893,
-    "eval_samples_per_second": 154.623,
-    "eval_steps_per_second": 4.909
+    "epoch": 9.98,
+    "eval_accuracy": 0.9828042328042328,
+    "eval_loss": 0.06841638684272766,
+    "eval_runtime": 5.801,
+    "eval_samples_per_second": 130.321,
+    "eval_steps_per_second": 4.137
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.98,
-    "total_flos": 5.346226696020296e+17,
-    "train_loss": 0.37407838611891775,
-    "train_runtime": 386.1594,
-    "train_samples_per_second": 55.935,
-    "train_steps_per_second": 0.427
+    "epoch": 9.98,
+    "total_flos": 1.0716320475504968e+18,
+    "train_loss": 0.24182906033414783,
+    "train_runtime": 827.7733,
+    "train_samples_per_second": 52.188,
+    "train_steps_per_second": 0.399
 }

trainer_state.json

@@ -1,166 +1,313 @@
 {
-  "best_metric": 0.9775132275132276,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-eurosat/checkpoint-165",
-  "epoch": 4.977777777777778,
-  "global_step": 165,
+  "best_metric": 0.9828042328042328,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-eurosat/checkpoint-297",
+  "epoch": 9.977777777777778,
+  "global_step": 330,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.3,
-      "learning_rate": 2.9411764705882354e-05,
-      "loss": 2.2557,
+      "learning_rate": 1.5151515151515153e-05,
+      "loss": 2.3574,
       "step": 10
     },
     {
       "epoch": 0.59,
-      "learning_rate": 4.8986486486486486e-05,
-      "loss": 1.5384,
+      "learning_rate": 3.0303030303030306e-05,
+      "loss": 1.981,
       "step": 20
     },
     {
       "epoch": 0.89,
-      "learning_rate": 4.560810810810811e-05,
-      "loss": 0.6233,
+      "learning_rate": 4.545454545454546e-05,
+      "loss": 1.2075,
       "step": 30
     },
     {
       "epoch": 0.98,
-      "eval_accuracy": 0.91005291005291,
-      "eval_loss": 0.30006203055381775,
-      "eval_runtime": 4.7638,
-      "eval_samples_per_second": 158.696,
-      "eval_steps_per_second": 5.038,
+      "eval_accuracy": 0.8518518518518519,
+      "eval_loss": 0.5666351914405823,
+      "eval_runtime": 4.8236,
+      "eval_samples_per_second": 156.731,
+      "eval_steps_per_second": 4.976,
       "step": 33
     },
     {
       "epoch": 1.21,
-      "learning_rate": 4.222972972972973e-05,
-      "loss": 0.3829,
+      "learning_rate": 4.882154882154882e-05,
+      "loss": 0.5066,
       "step": 40
     },
     {
       "epoch": 1.5,
-      "learning_rate": 3.885135135135135e-05,
-      "loss": 0.2118,
+      "learning_rate": 4.713804713804714e-05,
+      "loss": 0.2643,
       "step": 50
     },
     {
       "epoch": 1.8,
-      "learning_rate": 3.547297297297297e-05,
-      "loss": 0.1958,
+      "learning_rate": 4.545454545454546e-05,
+      "loss": 0.2022,
       "step": 60
     },
     {
       "epoch": 1.98,
-      "eval_accuracy": 0.9642857142857143,
-      "eval_loss": 0.12865345180034637,
-      "eval_runtime": 4.9076,
-      "eval_samples_per_second": 154.047,
-      "eval_steps_per_second": 4.89,
+      "eval_accuracy": 0.9126984126984127,
+      "eval_loss": 0.252269446849823,
+      "eval_runtime": 4.9657,
+      "eval_samples_per_second": 152.245,
+      "eval_steps_per_second": 4.833,
       "step": 66
     },
     {
       "epoch": 2.12,
-      "learning_rate": 3.20945945945946e-05,
-      "loss": 0.1691,
+      "learning_rate": 4.3771043771043774e-05,
+      "loss": 0.1839,
       "step": 70
     },
     {
       "epoch": 2.41,
-      "learning_rate": 2.8716216216216217e-05,
-      "loss": 0.1239,
+      "learning_rate": 4.208754208754209e-05,
+      "loss": 0.1368,
       "step": 80
     },
     {
       "epoch": 2.71,
-      "learning_rate": 2.533783783783784e-05,
-      "loss": 0.1212,
+      "learning_rate": 4.0404040404040405e-05,
+      "loss": 0.1206,
       "step": 90
     },
     {
       "epoch": 2.98,
-      "eval_accuracy": 0.9708994708994709,
-      "eval_loss": 0.11089407652616501,
-      "eval_runtime": 4.7941,
-      "eval_samples_per_second": 157.693,
-      "eval_steps_per_second": 5.006,
+      "eval_accuracy": 0.9497354497354498,
+      "eval_loss": 0.15757465362548828,
+      "eval_runtime": 4.8701,
+      "eval_samples_per_second": 155.233,
+      "eval_steps_per_second": 4.928,
       "step": 99
     },
     {
       "epoch": 3.03,
-      "learning_rate": 2.195945945945946e-05,
-      "loss": 0.1071,
+      "learning_rate": 3.872053872053872e-05,
+      "loss": 0.105,
       "step": 100
     },
     {
       "epoch": 3.33,
-      "learning_rate": 1.8581081081081082e-05,
-      "loss": 0.0814,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.0922,
       "step": 110
     },
     {
       "epoch": 3.62,
-      "learning_rate": 1.5202702702702704e-05,
-      "loss": 0.0962,
+      "learning_rate": 3.535353535353535e-05,
+      "loss": 0.0925,
       "step": 120
     },
     {
       "epoch": 3.92,
-      "learning_rate": 1.1824324324324325e-05,
-      "loss": 0.0734,
+      "learning_rate": 3.3670033670033675e-05,
+      "loss": 0.0897,
       "step": 130
     },
     {
       "epoch": 3.98,
-      "eval_accuracy": 0.9642857142857143,
-      "eval_loss": 0.11790694296360016,
-      "eval_runtime": 4.8978,
-      "eval_samples_per_second": 154.355,
-      "eval_steps_per_second": 4.9,
+      "eval_accuracy": 0.9563492063492064,
+      "eval_loss": 0.142096146941185,
+      "eval_runtime": 6.3572,
+      "eval_samples_per_second": 118.921,
+      "eval_steps_per_second": 3.775,
       "step": 132
     },
     {
       "epoch": 4.24,
-      "learning_rate": 8.445945945945946e-06,
-      "loss": 0.0679,
+      "learning_rate": 3.198653198653199e-05,
+      "loss": 0.0704,
       "step": 140
     },
     {
       "epoch": 4.53,
-      "learning_rate": 5.067567567567568e-06,
-      "loss": 0.0516,
+      "learning_rate": 3.0303030303030306e-05,
+      "loss": 0.0511,
       "step": 150
     },
     {
       "epoch": 4.83,
-      "learning_rate": 1.6891891891891894e-06,
-      "loss": 0.0457,
+      "learning_rate": 2.8619528619528618e-05,
+      "loss": 0.0564,
       "step": 160
     },
     {
       "epoch": 4.98,
-      "eval_accuracy": 0.9775132275132276,
-      "eval_loss": 0.09184322506189346,
-      "eval_runtime": 4.7828,
-      "eval_samples_per_second": 158.066,
-      "eval_steps_per_second": 5.018,
+      "eval_accuracy": 0.9656084656084656,
+      "eval_loss": 0.11138752847909927,
+      "eval_runtime": 5.5251,
+      "eval_samples_per_second": 136.831,
+      "eval_steps_per_second": 4.344,
       "step": 165
     },
     {
-      "epoch": 4.98,
-      "step": 165,
-      "total_flos": 5.346226696020296e+17,
-      "train_loss": 0.37407838611891775,
-      "train_runtime": 386.1594,
-      "train_samples_per_second": 55.935,
-      "train_steps_per_second": 0.427
+      "epoch": 5.15,
+      "learning_rate": 2.6936026936026937e-05,
+      "loss": 0.0559,
+      "step": 170
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.5252525252525256e-05,
+      "loss": 0.0468,
+      "step": 180
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 2.356902356902357e-05,
+      "loss": 0.0475,
+      "step": 190
+    },
+    {
+      "epoch": 5.98,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.06781856715679169,
+      "eval_runtime": 5.3143,
+      "eval_samples_per_second": 142.259,
+      "eval_steps_per_second": 4.516,
+      "step": 198
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 2.1885521885521887e-05,
+      "loss": 0.0514,
+      "step": 200
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 2.0202020202020203e-05,
+      "loss": 0.0237,
+      "step": 210
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.0255,
+      "step": 220
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 1.6835016835016837e-05,
+      "loss": 0.0332,
+      "step": 230
+    },
+    {
+      "epoch": 6.98,
+      "eval_accuracy": 0.9775132275132276,
+      "eval_loss": 0.08194286376237869,
+      "eval_runtime": 5.0836,
+      "eval_samples_per_second": 148.714,
+      "eval_steps_per_second": 4.721,
+      "step": 231
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.5151515151515153e-05,
+      "loss": 0.0263,
+      "step": 240
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.3468013468013468e-05,
+      "loss": 0.0193,
+      "step": 250
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 1.1784511784511786e-05,
+      "loss": 0.0234,
+      "step": 260
+    },
+    {
+      "epoch": 7.98,
+      "eval_accuracy": 0.9801587301587301,
+      "eval_loss": 0.06791768968105316,
+      "eval_runtime": 5.0567,
+      "eval_samples_per_second": 149.505,
+      "eval_steps_per_second": 4.746,
+      "step": 264
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.0101010101010101e-05,
+      "loss": 0.0193,
+      "step": 270
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 8.417508417508419e-06,
+      "loss": 0.0151,
+      "step": 280
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 6.734006734006734e-06,
+      "loss": 0.0126,
+      "step": 290
+    },
+    {
+      "epoch": 8.98,
+      "eval_accuracy": 0.9828042328042328,
+      "eval_loss": 0.06841638684272766,
+      "eval_runtime": 4.8183,
+      "eval_samples_per_second": 156.902,
+      "eval_steps_per_second": 4.981,
+      "step": 297
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 5.050505050505051e-06,
+      "loss": 0.0096,
+      "step": 300
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 3.367003367003367e-06,
+      "loss": 0.0091,
+      "step": 310
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.6835016835016836e-06,
+      "loss": 0.0136,
+      "step": 320
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 0.0,
+      "loss": 0.0306,
+      "step": 330
+    },
+    {
+      "epoch": 9.98,
+      "eval_accuracy": 0.9814814814814815,
+      "eval_loss": 0.07185560464859009,
+      "eval_runtime": 6.808,
+      "eval_samples_per_second": 111.046,
+      "eval_steps_per_second": 3.525,
+      "step": 330
+    },
+    {
+      "epoch": 9.98,
+      "step": 330,
+      "total_flos": 1.0716320475504968e+18,
+      "train_loss": 0.24182906033414783,
+      "train_runtime": 827.7733,
+      "train_samples_per_second": 52.188,
+      "train_steps_per_second": 0.399
     }
   ],
-  "max_steps": 165,
-  "num_train_epochs": 5,
-  "total_flos": 5.346226696020296e+17,
+  "max_steps": 330,
+  "num_train_epochs": 10,
+  "total_flos": 1.0716320475504968e+18,
   "trial_name": null,
   "trial_params": null
 }