Brain MRI 🧠 fine tune - 14 epochs.

README.md

@@ -3,6 +3,7 @@ library_name: transformers
 license: apache-2.0
 base_model: google/vit-base-patch16-224
 tags:
+- image-classification
 - generated_from_trainer
 metrics:
 - accuracy
@@ -16,10 +17,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # vit-base-brain-mri
 
-This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset.
+This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the BrainMRI dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5857
-- Accuracy: 0.8223
+- Loss: 1.1297
+- Accuracy: 0.5685
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 10.0,
-    "eval_accuracy": 0.5456852791878173,
-    "eval_loss": 1.1441423892974854,
-    "eval_runtime": 3.3533,
-    "eval_samples_per_second": 117.494,
-    "eval_steps_per_second": 14.91,
-    "total_flos": 1.779247974652969e+18,
-    "train_loss": 0.3026825189590454,
-    "train_runtime": 130.7944,
-    "train_samples_per_second": 175.543,
-    "train_steps_per_second": 5.505
+    "epoch": 14.0,
+    "eval_accuracy": 0.5685279187817259,
+    "eval_loss": 1.129746437072754,
+    "eval_runtime": 3.2309,
+    "eval_samples_per_second": 121.946,
+    "eval_steps_per_second": 15.475,
+    "total_flos": 2.4909471645141565e+18,
+    "train_loss": 0.15913245034596277,
+    "train_runtime": 101.1425,
+    "train_samples_per_second": 317.809,
+    "train_steps_per_second": 9.966
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 10.0,
-    "eval_accuracy": 0.5456852791878173,
-    "eval_loss": 1.1441423892974854,
-    "eval_runtime": 3.3533,
-    "eval_samples_per_second": 117.494,
-    "eval_steps_per_second": 14.91
+    "epoch": 14.0,
+    "eval_accuracy": 0.5685279187817259,
+    "eval_loss": 1.129746437072754,
+    "eval_runtime": 3.2309,
+    "eval_samples_per_second": 121.946,
+    "eval_steps_per_second": 15.475
 }

runs/Aug25_15-13-07_DESKTOP-4SGMSGR/events.out.tfevents.1724588107.DESKTOP-4SGMSGR.40716.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4336bd93a62feb8a9f868135ffff22ffd2bf11a93ac44a60b9681a391fbdaabc
+size 363

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 10.0,
-    "total_flos": 1.779247974652969e+18,
-    "train_loss": 0.3026825189590454,
-    "train_runtime": 130.7944,
-    "train_samples_per_second": 175.543,
-    "train_steps_per_second": 5.505
+    "epoch": 14.0,
+    "total_flos": 2.4909471645141565e+18,
+    "train_loss": 0.15913245034596277,
+    "train_runtime": 101.1425,
+    "train_samples_per_second": 317.809,
+    "train_steps_per_second": 9.966
 }

trainer_state.json

@@ -1,9 +1,9 @@
 {
-  "best_metric": 0.6089015603065491,
-  "best_model_checkpoint": "./vit-base-brain-mri\\checkpoint-648",
-  "epoch": 10.0,
+  "best_metric": 0.5857352614402771,
+  "best_model_checkpoint": "./vit-base-brain-mri\\checkpoint-1008",
+  "epoch": 14.0,
   "eval_steps": 500,
-  "global_step": 720,
+  "global_step": 1008,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -148,19 +148,76 @@
       "step": 720
     },
     {
-      "epoch": 10.0,
-      "step": 720,
-      "total_flos": 1.779247974652969e+18,
-      "train_loss": 0.3026825189590454,
-      "train_runtime": 130.7944,
-      "train_samples_per_second": 175.543,
-      "train_steps_per_second": 5.505
+      "epoch": 11.0,
+      "eval_accuracy": 0.7839721254355401,
+      "eval_loss": 0.6184366941452026,
+      "eval_runtime": 4.8681,
+      "eval_samples_per_second": 117.91,
+      "eval_steps_per_second": 14.79,
+      "step": 792
+    },
+    {
+      "epoch": 11.11111111111111,
+      "grad_norm": 2.4091649055480957,
+      "learning_rate": 6.190476190476189e-05,
+      "loss": 0.5733,
+      "step": 800
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.794425087108014,
+      "eval_loss": 0.6006320714950562,
+      "eval_runtime": 4.8997,
+      "eval_samples_per_second": 117.151,
+      "eval_steps_per_second": 14.695,
+      "step": 864
+    },
+    {
+      "epoch": 12.5,
+      "grad_norm": 2.6921584606170654,
+      "learning_rate": 3.214285714285714e-05,
+      "loss": 0.5555,
+      "step": 900
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.8013937282229965,
+      "eval_loss": 0.5897805690765381,
+      "eval_runtime": 4.8766,
+      "eval_samples_per_second": 117.705,
+      "eval_steps_per_second": 14.764,
+      "step": 936
+    },
+    {
+      "epoch": 13.88888888888889,
+      "grad_norm": 1.2396776676177979,
+      "learning_rate": 2.3809523809523808e-06,
+      "loss": 0.5481,
+      "step": 1000
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.8222996515679443,
+      "eval_loss": 0.5857352614402771,
+      "eval_runtime": 4.9368,
+      "eval_samples_per_second": 116.27,
+      "eval_steps_per_second": 14.584,
+      "step": 1008
+    },
+    {
+      "epoch": 14.0,
+      "step": 1008,
+      "total_flos": 2.4909471645141565e+18,
+      "train_loss": 0.15913245034596277,
+      "train_runtime": 101.1425,
+      "train_samples_per_second": 317.809,
+      "train_steps_per_second": 9.966
     }
   ],
   "logging_steps": 100,
-  "max_steps": 720,
+  "max_steps": 1008,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 14,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -174,7 +231,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.779247974652969e+18,
+  "total_flos": 2.4909471645141565e+18,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null