End of training

README.md

@@ -12,12 +12,12 @@ probably proofread and complete it, then remove this comment. -->
 
 # shbA/food_classifier
 
-This model is a fine-tuned version of [facebook/dino-vitb16](https://huggingface.co/facebook/dino-vitb16) on an unknown dataset.
+This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Train Loss: 0.4493
-- Validation Loss: 0.4618
-- Train Accuracy: 0.8043
-- Epoch: 6
+- Train Loss: 0.5020
+- Validation Loss: 0.4878
+- Train Accuracy: 0.8125
+- Epoch: 9
 
 ## Model description
 
@@ -36,20 +36,23 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 3e-05, 'decay_steps': 1288, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
+- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 3e-05, 'decay_steps': 1920, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.05}
 - training_precision: float32
 
 ### Training results
 
 | Train Loss | Validation Loss | Train Accuracy | Epoch |
 |:----------:|:---------------:|:--------------:|:-----:|
-| 1.2277     | 0.8208          | 0.6087         | 0     |
-| 0.7967     | 0.4934          | 0.8261         | 1     |
-| 0.5488     | 0.5541          | 0.7391         | 2     |
-| 0.6664     | 1.0464          | 0.6304         | 3     |
-| 0.7975     | 1.1285          | 0.5652         | 4     |
-| 0.6134     | 0.8258          | 0.6957         | 5     |
-| 0.4493     | 0.4618          | 0.8043         | 6     |
+| 1.0700     | 0.9871          | 0.5208         | 0     |
+| 0.9347     | 0.7458          | 0.7917         | 1     |
+| 0.8570     | 0.7546          | 0.6875         | 2     |
+| 0.7179     | 0.5846          | 0.75           | 3     |
+| 0.6272     | 0.6125          | 0.75           | 4     |
+| 0.5631     | 0.4566          | 0.8125         | 5     |
+| 0.5580     | 0.4655          | 0.7917         | 6     |
+| 0.5796     | 0.4462          | 0.8333         | 7     |
+| 0.4934     | 0.4899          | 0.8542         | 8     |
+| 0.5020     | 0.4878          | 0.8125         | 9     |
 
 
 ### Framework versions

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/dino-vitb16",
+  "_name_or_path": "google/vit-base-patch16-224-in21k",
   "architectures": [
     "ViTForImageClassification"
   ],
@@ -28,6 +28,5 @@
   "num_hidden_layers": 12,
   "patch_size": 16,
   "qkv_bias": true,
-  "torch_dtype": "float32",
   "transformers_version": "4.29.1"
 }

preprocessor_config.json

@@ -3,15 +3,15 @@
   "do_rescale": true,
   "do_resize": true,
   "image_mean": [
-    0.485,
-    0.456,
-    0.406
+    0.5,
+    0.5,
+    0.5
   ],
   "image_processor_type": "ViTImageProcessor",
   "image_std": [
-    0.229,
-    0.224,
-    0.225
+    0.5,
+    0.5,
+    0.5
   ],
   "resample": 2,
   "rescale_factor": 0.00392156862745098,

tf_model.h5

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:831d89bffc0c679feffdc68af98122c90fc885a2d208843c43849c10ee64a58b
+oid sha256:c762ce5987515e2a1157e2255a191beade55bb867443d9d1c208c9bba23b3fcd
 size 343472824