Training in progress, epoch 1

all_results.json

@@ -1,13 +1,8 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.9911111111111112,
-    "eval_loss": 0.02681696228682995,
-    "eval_runtime": 1.8717,
-    "eval_samples_per_second": 120.214,
-    "eval_steps_per_second": 4.274,
-    "total_flos": 1.5048656676458496e+17,
-    "train_loss": 0.5111757349222898,
-    "train_runtime": 85.6924,
-    "train_samples_per_second": 470.987,
-    "train_steps_per_second": 3.734
+    "epoch": 2.873239436619718,
+    "total_flos": 1.608272360368128e+17,
+    "train_loss": 0.2495375123678469,
+    "train_runtime": 94.7041,
+    "train_samples_per_second": 71.053,
+    "train_steps_per_second": 0.539
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d4daa43b61af68d55247d00ac988cc41108b008bcf2ff4d64cef35c4bf23be49
+oid sha256:95e437a98d9d912f824b5169bedf1fa4934dde54bae7343a6cc210d2be8c293a
 size 110348984

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "total_flos": 1.5048656676458496e+17,
-    "train_loss": 0.5111757349222898,
-    "train_runtime": 85.6924,
-    "train_samples_per_second": 470.987,
-    "train_steps_per_second": 3.734
+    "epoch": 2.873239436619718,
+    "total_flos": 1.608272360368128e+17,
+    "train_loss": 0.2495375123678469,
+    "train_runtime": 94.7041,
+    "train_samples_per_second": 71.053,
+    "train_steps_per_second": 0.539
 }

trainer_state.json

@@ -1,93 +1,77 @@
 {
-  "best_metric": 0.9911111111111112,
-  "best_model_checkpoint": "/kaggle/working/swin-brain-modality-classification/checkpoint-48",
-  "epoch": 3.0,
+  "best_metric": 0.6410256410256411,
+  "best_model_checkpoint": "/kaggle/working/swin-brain-modality-classification/checkpoint-35",
+  "epoch": 2.873239436619718,
   "eval_steps": 500,
-  "global_step": 48,
+  "global_step": 51,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.625,
-      "grad_norm": 9.655741691589355,
-      "learning_rate": 1.5625e-05,
-      "loss": 1.364,
-      "step": 10
+      "epoch": 0.9577464788732394,
+      "grad_norm": 4.190108776092529,
+      "learning_rate": 3.777777777777778e-05,
+      "loss": 0.6378,
+      "step": 17
     },
     {
-      "epoch": 1.0,
-      "eval_accuracy": 0.8266666666666667,
-      "eval_loss": 0.6507545113563538,
-      "eval_runtime": 1.8483,
-      "eval_samples_per_second": 121.736,
-      "eval_steps_per_second": 4.328,
-      "step": 16
+      "epoch": 0.9577464788732394,
+      "eval_accuracy": 0.6239316239316239,
+      "eval_loss": 3.3409788608551025,
+      "eval_runtime": 3.7527,
+      "eval_samples_per_second": 124.711,
+      "eval_steps_per_second": 3.997,
+      "step": 17
     },
     {
-      "epoch": 1.25,
-      "grad_norm": 6.807246685028076,
-      "learning_rate": 3.125e-05,
-      "loss": 0.7521,
-      "step": 20
+      "epoch": 1.971830985915493,
+      "grad_norm": 3.4180703163146973,
+      "learning_rate": 1.777777777777778e-05,
+      "loss": 0.0669,
+      "step": 35
     },
     {
-      "epoch": 1.875,
-      "grad_norm": 5.148420333862305,
-      "learning_rate": 4.6875e-05,
-      "loss": 0.2053,
-      "step": 30
+      "epoch": 1.971830985915493,
+      "eval_accuracy": 0.6410256410256411,
+      "eval_loss": 4.32125186920166,
+      "eval_runtime": 3.6088,
+      "eval_samples_per_second": 129.681,
+      "eval_steps_per_second": 4.156,
+      "step": 35
     },
     {
-      "epoch": 2.0,
-      "eval_accuracy": 0.9555555555555556,
-      "eval_loss": 0.08017182350158691,
-      "eval_runtime": 1.8504,
-      "eval_samples_per_second": 121.592,
-      "eval_steps_per_second": 4.323,
-      "step": 32
+      "epoch": 2.873239436619718,
+      "grad_norm": 4.94912052154541,
+      "learning_rate": 0.0,
+      "loss": 0.0425,
+      "step": 51
     },
     {
-      "epoch": 2.5,
-      "grad_norm": 9.427779197692871,
-      "learning_rate": 4.8611111111111115e-05,
-      "loss": 0.0841,
-      "step": 40
+      "epoch": 2.873239436619718,
+      "eval_accuracy": 0.6410256410256411,
+      "eval_loss": 4.431879997253418,
+      "eval_runtime": 3.8435,
+      "eval_samples_per_second": 121.765,
+      "eval_steps_per_second": 3.903,
+      "step": 51
     },
     {
-      "epoch": 3.0,
-      "eval_accuracy": 0.9911111111111112,
-      "eval_loss": 0.02681696228682995,
-      "eval_runtime": 1.886,
-      "eval_samples_per_second": 119.3,
-      "eval_steps_per_second": 4.242,
-      "step": 48
-    },
-    {
-      "epoch": 3.0,
-      "step": 48,
-      "total_flos": 1.5048656676458496e+17,
-      "train_loss": 0.5111757349222898,
-      "train_runtime": 85.6924,
-      "train_samples_per_second": 470.987,
-      "train_steps_per_second": 3.734
+      "epoch": 2.873239436619718,
+      "step": 51,
+      "total_flos": 1.608272360368128e+17,
+      "train_loss": 0.2495375123678469,
+      "train_runtime": 94.7041,
+      "train_samples_per_second": 71.053,
+      "train_steps_per_second": 0.539
     }
   ],
-  "logging_steps": 10,
-  "max_steps": 320,
+  "logging_steps": 1,
+  "max_steps": 51,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 20,
+  "num_train_epochs": 3,
   "save_steps": 500,
   "stateful_callbacks": {
-    "CustomEarlyStoppingCallback": {
-      "args": {
-        "early_stopping_patience": 1,
-        "early_stopping_threshold": 0.0
-      },
-      "attributes": {
-        "early_stopping_patience_counter": 0
-      }
-    },
     "TrainerControl": {
       "args": {
         "should_epoch_stop": false,
@@ -99,7 +83,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.5048656676458496e+17,
+  "total_flos": 1.608272360368128e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null