template

app.py

@@ -0,0 +1,63 @@
+import gradio as gr
+from gradio_imageslider import ImageSlider
+from loadimg import load_img
+import spaces
+from transformers import AutoModelForImageSegmentation
+import torch
+from torchvision import transforms
+
+torch.set_float32_matmul_precision(["high", "highest"][0])
+
+birefnet = AutoModelForImageSegmentation.from_pretrained(
+    "ZhengPeng7/BiRefNet", trust_remote_code=True
+)
+birefnet.to("cuda")
+transform_image = transforms.Compose(
+    [
+        transforms.Resize((1024, 1024)),
+        transforms.ToTensor(),
+        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
+    ]
+)
+
+def fn(vid):
+    # TODO 
+    # loop over video and extract images and process each one
+    im = load_img(vid, output_type="pil")
+    im = im.convert("RGB")
+    image = process(im)
+    return image
+
+@spaces.GPU
+def process(image):
+    image_size = image.size
+    input_images = transform_image(image).unsqueeze(0).to("cuda")
+    # Prediction
+    with torch.no_grad():
+        preds = birefnet(input_images)[-1].sigmoid().cpu()
+    pred = preds[0].squeeze()
+    pred_pil = transforms.ToPILImage()(pred)
+    mask = pred_pil.resize(image_size)
+    image.putalpha(mask)
+    return image
+  
+def process_file(f):
+    name_path = f.rsplit(".",1)[0]+".png"
+    im = load_img(f, output_type="pil")
+    im = im.convert("RGB")
+    transparent = process(im)
+    transparent.save(name_path)
+    return name_path
+
+in_video = gr.Video(label="birefnet")
+out_video = gr.Video()
+
+
+url = "https://hips.hearstapps.com/hmg-prod/images/gettyimages-1229892983-square.jpg"
+demo = gr.Interface(
+    fn, inputs=in_video, outputs=out_video, api_name="image"
+)
+
+
+if __name__ == "__main__":
+    demo.launch(show_error=True)

requirements.txt

@@ -0,0 +1,16 @@
+torch
+accelerate
+opencv-python
+spaces
+pillow
+numpy
+timm
+kornia
+prettytable
+typing
+scikit-image
+huggingface_hub
+transformers>=4.39.1
+gradio
+gradio_imageslider
+loadimg>=0.1.1