Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import CLIPProcessor, CLIPModel
|
3 |
+
from PIL import Image
|
4 |
+
import torch
|
5 |
+
|
6 |
+
# Load CLIP model and processor
|
7 |
+
model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
|
8 |
+
processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
|
9 |
+
|
10 |
+
# Define a list of target words for the game
|
11 |
+
words = ["cat", "car", "tree", "house", "dog"]
|
12 |
+
|
13 |
+
|
14 |
+
text_inputs = processor(text=words, return_tensors="pt", padding=True)
|
15 |
+
with torch.no_grad():
|
16 |
+
text_features = model.get_text_features(**text_inputs)
|
17 |
+
|
18 |
+
|
19 |
+
def guess_drawing(drawing):
|
20 |
+
image = Image.fromarray(drawing) # Convert drawing to PIL image
|
21 |
+
image_inputs = processor(images=image, return_tensors="pt")
|
22 |
+
with torch.no_grad():
|
23 |
+
image_features = model.get_image_features(**image_inputs)
|
24 |
+
|
25 |
+
# Calculate cosine similarity with each word
|
26 |
+
similarity = torch.nn.functional.cosine_similarity(image_features, text_features)
|
27 |
+
best_match = words[similarity.argmax().item()]
|
28 |
+
return f"AI's guess: {best_match}"
|
29 |
+
|
30 |
+
|
31 |
+
interface = gr.Interface(
|
32 |
+
fn=guess_drawing,
|
33 |
+
inputs=gr.Sketchpad(),
|
34 |
+
outputs="text",
|
35 |
+
live=True,
|
36 |
+
description="Draw something and see if the AI can guess it!"
|
37 |
+
)
|
38 |
+
|
39 |
+
interface.launch()
|
40 |
+
|