Update app.py
Browse files
app.py
CHANGED
@@ -8,8 +8,9 @@ model, preprocess = clip.load("ViT-B/32", device=device)
|
|
8 |
|
9 |
def clip(image, text):
|
10 |
|
|
|
11 |
image = preprocess(image).unsqueeze(0).to(device)
|
12 |
-
text = clip.tokenize(
|
13 |
|
14 |
with torch.no_grad():
|
15 |
image_features = model.encode_image(image)
|
@@ -18,7 +19,7 @@ def clip(image, text):
|
|
18 |
logits_per_image, logits_per_text = model(image, text)
|
19 |
probs = logits_per_image.softmax(dim=-1).cpu().numpy()
|
20 |
|
21 |
-
return probs
|
22 |
|
23 |
demo = gr.Interface(fn=clip, inputs=["text", "image"], outputs="text")
|
24 |
demo.launch()
|
|
|
8 |
|
9 |
def clip(image, text):
|
10 |
|
11 |
+
text = text.split(",")
|
12 |
image = preprocess(image).unsqueeze(0).to(device)
|
13 |
+
text = clip.tokenize(text).to(device)
|
14 |
|
15 |
with torch.no_grad():
|
16 |
image_features = model.encode_image(image)
|
|
|
19 |
logits_per_image, logits_per_text = model(image, text)
|
20 |
probs = logits_per_image.softmax(dim=-1).cpu().numpy()
|
21 |
|
22 |
+
return probs
|
23 |
|
24 |
demo = gr.Interface(fn=clip, inputs=["text", "image"], outputs="text")
|
25 |
demo.launch()
|