End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 4.0,
-    "eval_accuracy": 1.0,
-    "eval_loss": 0.19110850989818573,
-    "eval_runtime": 0.3814,
-    "eval_samples_per_second": 165.197,
-    "eval_steps_per_second": 5.244,
-    "total_flos": 5.637495822800486e+16,
-    "train_loss": 0.6830070416132609,
-    "train_runtime": 28.7573,
-    "train_samples_per_second": 98.584,
-    "train_steps_per_second": 0.695
+    "epoch": 9.0,
+    "eval_accuracy": 0.9851851851851852,
+    "eval_loss": 0.063043974339962,
+    "eval_runtime": 0.7297,
+    "eval_samples_per_second": 185.014,
+    "eval_steps_per_second": 6.852,
+    "total_flos": 1.4093739557001216e+17,
+    "train_loss": 0.3985253996319241,
+    "train_runtime": 73.599,
+    "train_samples_per_second": 171.198,
+    "train_steps_per_second": 1.359
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.0,
-    "eval_accuracy": 1.0,
-    "eval_loss": 0.19110850989818573,
-    "eval_runtime": 0.3814,
-    "eval_samples_per_second": 165.197,
-    "eval_steps_per_second": 5.244
+    "epoch": 9.0,
+    "eval_accuracy": 0.9851851851851852,
+    "eval_loss": 0.063043974339962,
+    "eval_runtime": 0.7297,
+    "eval_samples_per_second": 185.014,
+    "eval_steps_per_second": 6.852
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.0,
-    "total_flos": 5.637495822800486e+16,
-    "train_loss": 0.6830070416132609,
-    "train_runtime": 28.7573,
-    "train_samples_per_second": 98.584,
-    "train_steps_per_second": 0.695
+    "epoch": 9.0,
+    "total_flos": 1.4093739557001216e+17,
+    "train_loss": 0.3985253996319241,
+    "train_runtime": 73.599,
+    "train_samples_per_second": 171.198,
+    "train_steps_per_second": 1.359
 }

trainer_state.json

@@ -1,70 +1,171 @@
 {
-  "best_metric": 1.0,
-  "best_model_checkpoint": "/kaggle/working/swin-brain-plane-classification/checkpoint-18",
-  "epoch": 4.0,
+  "best_metric": 0.9851851851851852,
+  "best_model_checkpoint": "/kaggle/working/swin-brain-plane-classification/checkpoint-45",
+  "epoch": 9.0,
   "eval_steps": 500,
-  "global_step": 18,
+  "global_step": 45,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.8888888888888888,
-      "eval_accuracy": 0.9047619047619048,
-      "eval_loss": 0.8267417550086975,
-      "eval_runtime": 0.3659,
-      "eval_samples_per_second": 172.198,
-      "eval_steps_per_second": 5.467,
-      "step": 4
+      "epoch": 1.0,
+      "grad_norm": 5.971311092376709,
+      "learning_rate": 2.5e-05,
+      "loss": 1.0732,
+      "step": 5
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 0.9284645318984985,
+      "eval_runtime": 0.7118,
+      "eval_samples_per_second": 189.658,
+      "eval_steps_per_second": 7.024,
+      "step": 5
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 10.538399696350098,
+      "learning_rate": 5e-05,
+      "loss": 0.829,
+      "step": 10
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.9682539682539683,
-      "eval_loss": 0.45388203859329224,
-      "eval_runtime": 0.3836,
-      "eval_samples_per_second": 164.226,
-      "eval_steps_per_second": 5.214,
-      "step": 9
-    },
-    {
-      "epoch": 2.2222222222222223,
-      "grad_norm": 19.5561466217041,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.8656,
+      "eval_accuracy": 0.8592592592592593,
+      "eval_loss": 0.527397871017456,
+      "eval_runtime": 0.7119,
+      "eval_samples_per_second": 189.625,
+      "eval_steps_per_second": 7.023,
       "step": 10
     },
     {
-      "epoch": 2.888888888888889,
-      "eval_accuracy": 0.9682539682539683,
-      "eval_loss": 0.30353549122810364,
-      "eval_runtime": 0.368,
-      "eval_samples_per_second": 171.182,
-      "eval_steps_per_second": 5.434,
-      "step": 13
+      "epoch": 3.0,
+      "grad_norm": 6.285182476043701,
+      "learning_rate": 4.722222222222222e-05,
+      "loss": 0.5189,
+      "step": 15
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9333333333333333,
+      "eval_loss": 0.24730348587036133,
+      "eval_runtime": 0.6998,
+      "eval_samples_per_second": 192.917,
+      "eval_steps_per_second": 7.145,
+      "step": 15
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.19110850989818573,
-      "eval_runtime": 0.3651,
-      "eval_samples_per_second": 172.56,
-      "eval_steps_per_second": 5.478,
-      "step": 18
+      "grad_norm": 9.056550979614258,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.3517,
+      "step": 20
     },
     {
       "epoch": 4.0,
-      "step": 18,
-      "total_flos": 5.637495822800486e+16,
-      "train_loss": 0.6830070416132609,
-      "train_runtime": 28.7573,
-      "train_samples_per_second": 98.584,
-      "train_steps_per_second": 0.695
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.1404803842306137,
+      "eval_runtime": 0.7065,
+      "eval_samples_per_second": 191.084,
+      "eval_steps_per_second": 7.077,
+      "step": 20
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 7.212975025177002,
+      "learning_rate": 4.166666666666667e-05,
+      "loss": 0.2075,
+      "step": 25
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9333333333333333,
+      "eval_loss": 0.17972101271152496,
+      "eval_runtime": 0.703,
+      "eval_samples_per_second": 192.032,
+      "eval_steps_per_second": 7.112,
+      "step": 25
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 5.206702709197998,
+      "learning_rate": 3.888888888888889e-05,
+      "loss": 0.1871,
+      "step": 30
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9703703703703703,
+      "eval_loss": 0.09362433105707169,
+      "eval_runtime": 0.759,
+      "eval_samples_per_second": 177.874,
+      "eval_steps_per_second": 6.588,
+      "step": 30
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 5.035118579864502,
+      "learning_rate": 3.611111111111111e-05,
+      "loss": 0.1567,
+      "step": 35
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9703703703703703,
+      "eval_loss": 0.1113051250576973,
+      "eval_runtime": 0.7175,
+      "eval_samples_per_second": 188.151,
+      "eval_steps_per_second": 6.969,
+      "step": 35
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 4.850553035736084,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.1482,
+      "step": 40
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9777777777777777,
+      "eval_loss": 0.0790533721446991,
+      "eval_runtime": 0.7241,
+      "eval_samples_per_second": 186.438,
+      "eval_steps_per_second": 6.905,
+      "step": 40
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 4.634683132171631,
+      "learning_rate": 3.055555555555556e-05,
+      "loss": 0.1143,
+      "step": 45
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9851851851851852,
+      "eval_loss": 0.063043974339962,
+      "eval_runtime": 0.7126,
+      "eval_samples_per_second": 189.438,
+      "eval_steps_per_second": 7.016,
+      "step": 45
+    },
+    {
+      "epoch": 9.0,
+      "step": 45,
+      "total_flos": 1.4093739557001216e+17,
+      "train_loss": 0.3985253996319241,
+      "train_runtime": 73.599,
+      "train_samples_per_second": 171.198,
+      "train_steps_per_second": 1.359
     }
   ],
-  "logging_steps": 10,
-  "max_steps": 20,
+  "logging_steps": 500,
+  "max_steps": 100,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 20,
   "save_steps": 500,
   "stateful_callbacks": {
     "CustomEarlyStoppingCallback": {
@@ -87,7 +188,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.637495822800486e+16,
+  "total_flos": 1.4093739557001216e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null