Update README.md
Browse files
README.md
CHANGED
@@ -7,18 +7,18 @@ A fine-tuned version of ViT-base on a collected set of Pokémon images. You can
|
|
7 |
# Using the model
|
8 |
|
9 |
```python
|
10 |
-
from transformers import ViTForImageClassification,
|
11 |
from PIL import Image
|
12 |
import torch
|
13 |
|
14 |
# Loading in Model
|
15 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
16 |
model = ViTForImageClassification.from_pretrained( "imjeffhi/pokemon_classifier").to(device)
|
17 |
-
|
18 |
|
19 |
# Caling the model on a test image
|
20 |
img = Image.open('test.jpg')
|
21 |
-
extracted =
|
22 |
predicted_id = model(**extracted).logits.argmax(-1).item()
|
23 |
predicted_pokemon = model.config.id2label[predicted_id]
|
24 |
```
|
|
|
7 |
# Using the model
|
8 |
|
9 |
```python
|
10 |
+
from transformers import ViTForImageClassification, ViTImageProcessor
|
11 |
from PIL import Image
|
12 |
import torch
|
13 |
|
14 |
# Loading in Model
|
15 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
16 |
model = ViTForImageClassification.from_pretrained( "imjeffhi/pokemon_classifier").to(device)
|
17 |
+
image_processor = ViTImageProcessor.from_pretrained('imjeffhi/pokemon_classifier')
|
18 |
|
19 |
# Caling the model on a test image
|
20 |
img = Image.open('test.jpg')
|
21 |
+
extracted = image_processor(images=img, return_tensors='pt').to(device)
|
22 |
predicted_id = model(**extracted).logits.argmax(-1).item()
|
23 |
predicted_pokemon = model.config.id2label[predicted_id]
|
24 |
```
|