CLIPSeg / app.py
nielsr's picture
nielsr HF staff
Create app.py
a5fde5f
from transformers import CLIPSegProcessor, CLIPSegForImageSegmentation
import gradio as gr
import torch
import matplotlib.pyplot as plt
import cv2
processor = CLIPSegProcessor.from_pretrained("CIDAS/clipseg-rd64-refined")
model = CLIPSegForImageSegmentation.from_pretrained("CIDAS/clipseg-rd64-refined")
def process_image(image, prompts):
inputs = processor(text=prompts, images=[image] * len(prompts), padding="max_length", return_tensors="pt")
# predict
with torch.no_grad():
outputs = model(**inputs)
preds = outputs.logits.unsqueeze(1)
filename = f"mask.png"
plt.imsave(filename,torch.sigmoid(preds[1][0]))
img2 = cv2.imread(filename)
gray_image = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
(thresh, bw_image) = cv2.threshold(gray_image, 100, 255, cv2.THRESH_BINARY)
# fix color format
cv2.cvtColor(bw_image, cv2.COLOR_BGR2RGB)
return Image.fromarray(bw_image)
title = "Interactive demo: zero-shot image segmentation with CLIPSeg"
description = "Demo for using CLIPSeg, a CLIP-based model for zero- and one-shot image segmentation. To use it, simply upload an image and add a text to mask (identify in the image), or use one of the examples below and click 'submit'. Results will show up in a few seconds."
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2112.10003'>CLIPSeg: Image Segmentation Using Text and Image Prompts</a> | <a href='https://huggingface.co/docs/transformers/main/en/model_doc/clipseg'>HuggingFace docs</a></p>"
examples = [["a glass", "something to fill", "wood", "a jar"]]
interface = gr.Interface(fn=process_image,
inputs=[gr.Image(type="pil"), gr.Textbox(label="What do you want to identify (separated by comma)?")],
outputs=gr.Image(type="pil"),
title=title,
description=description,
article=article,
examples=examples)
interface.launch(debug=True)