Spaces:
Sleeping
Sleeping
shaktibiplab
commited on
Upload app.py
Browse files
app.py
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
from transformers import pipeline, AutoTokenizer
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
# Load tokenizer
|
6 |
+
tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
|
7 |
+
tokenizer.clean_up_tokenization_spaces = False # Explicitly set the parameter if needed
|
8 |
+
|
9 |
+
# Load CLIP model for zero-shot classification
|
10 |
+
clip_checkpoint = "openai/clip-vit-base-patch16"
|
11 |
+
clip_detector = pipeline(model=clip_checkpoint, task="zero-shot-image-classification")
|
12 |
+
|
13 |
+
# Postprocess the output from CLIP
|
14 |
+
def postprocess(output):
|
15 |
+
return {out["label"]: float(out["score"]) for out in output}
|
16 |
+
|
17 |
+
# Inference function for CLIP
|
18 |
+
def infer(image, candidate_labels):
|
19 |
+
candidate_labels = [label.lstrip(" ") for label in candidate_labels.split(",")]
|
20 |
+
clip_out = clip_detector(image, candidate_labels=candidate_labels)
|
21 |
+
return postprocess(clip_out)
|
22 |
+
|
23 |
+
# Gradio interface
|
24 |
+
with gr.Blocks() as app:
|
25 |
+
gr.Markdown("# Custom Classification")
|
26 |
+
with gr.Row():
|
27 |
+
with gr.Column():
|
28 |
+
image_input = gr.Image(type="pil")
|
29 |
+
text_input = gr.Textbox(label="Input a list of labels")
|
30 |
+
run_button = gr.Button("Run")
|
31 |
+
|
32 |
+
with gr.Column():
|
33 |
+
clip_output = gr.Label(label="Output", num_top_classes=3)
|
34 |
+
|
35 |
+
examples = [["image_8.webp", "girl, boy, lgbtq"]]
|
36 |
+
gr.Examples(
|
37 |
+
examples=examples,
|
38 |
+
inputs=[image_input, text_input],
|
39 |
+
outputs=[clip_output],
|
40 |
+
fn=infer,
|
41 |
+
cache_examples=True
|
42 |
+
)
|
43 |
+
|
44 |
+
run_button.click(fn=infer,
|
45 |
+
inputs=[image_input, text_input],
|
46 |
+
outputs=[clip_output])
|
47 |
+
|
48 |
+
app.launch()
|