Update app.py

app.py

@@ -11,18 +11,24 @@ import numpy as np
 import os
 import tempfile
 import uuid
-import schedule
 import time
-import shutil
+import threading
 
 torch.set_float32_matmul_precision("medium")
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
+# Load both BiRefNet models
 birefnet = AutoModelForImageSegmentation.from_pretrained(
  "ZhengPeng7/BiRefNet", trust_remote_code=True
 )
 birefnet.to(device)
+
+birefnet_lite = AutoModelForImageSegmentation.from_pretrained(
+ "ZhengPeng7/BiRefNet_lite", trust_remote_code=True
+)
+birefnet_lite.to(device)
+
 transform_image = transforms.Compose(
  [
  transforms.Resize((1024, 1024)),
@@ -32,8 +38,31 @@ transform_image = transforms.Compose(
 )
 
 
+# Function to delete files older than 10 minutes in the temp directory
+def cleanup_temp_files():
+ while True:
+ temp_dir = "temp"
+ if os.path.exists(temp_dir):
+ for filename in os.listdir(temp_dir):
+ filepath = os.path.join(temp_dir, filename)
+ if os.path.isfile(filepath):
+ file_age = time.time() - os.path.getmtime(filepath)
+ if file_age > 600: # 10 minutes in seconds
+ try:
+ os.remove(filepath)
+ print(f"Deleted temporary file: {filepath}")
+ except Exception as e:
+ print(f"Error deleting file {filepath}: {e}")
+ time.sleep(60) # Check every minute
+
+
+# Start the cleanup thread
+cleanup_thread = threading.Thread(target=cleanup_temp_files, daemon=True)
+cleanup_thread.start()
+
+
 @spaces.GPU
-def fn(vid, bg_type="Color", bg_image=None, bg_video=None, color="#00FF00", fps=0, video_handling="slow_down"):
+def fn(vid, bg_type="Color", bg_image=None, bg_video=None, color="#00FF00", fps=0, video_handling="slow_down", fast_mode=False):
  try:
  # Load the video using moviepy
  video = mp.VideoFileClip(vid)
@@ -68,20 +97,20 @@ def fn(vid, bg_type="Color", bg_image=None, bg_video=None, color="#00FF00", fps=
  for i, frame in enumerate(frames):
  pil_image = Image.fromarray(frame)
  if bg_type == "Color":
- processed_image = process(pil_image, color)
+ processed_image = process(pil_image, color, fast_mode)
  elif bg_type == "Image":
- processed_image = process(pil_image, bg_image)
+ processed_image = process(pil_image, bg_image, fast_mode)
  elif bg_type == "Video":
  if video_handling == "slow_down":
  background_frame = background_frames[bg_frame_index % len(background_frames)]
  bg_frame_index += 1
  background_image = Image.fromarray(background_frame)
- processed_image = process(pil_image, background_image)
+ processed_image = process(pil_image, background_image, fast_mode)
  else: # video_handling == "loop"
  background_frame = background_frames[bg_frame_index % len(background_frames)]
  bg_frame_index += 1
  background_image = Image.fromarray(background_frame)
- processed_image = process(pil_image, background_image)
+ processed_image = process(pil_image, background_image, fast_mode)
  else:
  processed_image = pil_image # Default to original image if no background is selected
 
@@ -111,13 +140,16 @@ def fn(vid, bg_type="Color", bg_image=None, bg_video=None, color="#00FF00", fps=
  yield None, f"Error processing video: {e}"
 
 
-
-def process(image, bg):
+def process(image, bg, fast_mode=False):
  image_size = image.size
  input_images = transform_image(image).unsqueeze(0).to("cuda")
+
+ # Select the model based on fast_mode
+ model = birefnet_lite if fast_mode else birefnet
+
  # Prediction
  with torch.no_grad():
- preds = birefnet(input_images)[-1].sigmoid().cpu()
+ preds = model(input_images)[-1].sigmoid().cpu()
  pred = preds[0].squeeze()
  pred_pil = transforms.ToPILImage()(pred)
  mask = pred_pil.resize(image_size)
@@ -135,18 +167,6 @@ def process(image, bg):
 
  return image
 
-def clear_temp_directory():
- temp_dir = "temp"
- for filename in os.listdir(temp_dir):
- file_path = os.path.join(temp_dir, filename)
- try:
- if os.path.isfile(file_path) or os.path.islink(file_path):
- os.unlink(file_path)
- elif os.path.isdir(file_path):
- shutil.rmtree(file_path)
- except Exception as e:
- print('Failed to delete %s. Reason: %s' % (file_path, e)) # Keep this print statement for debugging purposes
-
 
 with gr.Blocks(theme=gr.themes.Ocean()) as demo:
  gr.Markdown("# Video Background Remover & Changer\n### You can replace image background with any color, image or video.\nNOTE: As this Space is running on ZERO GPU it has limit. It can handle approx 200frmaes at once. So, if you have big video than use small chunks or Duplicate this space.")
@@ -170,6 +190,8 @@ with gr.Blocks(theme=gr.themes.Ocean()) as demo:
  bg_video = gr.Video(label="Background Video", visible=False, interactive=True)
  with gr.Column(visible=False) as video_handling_options:
  video_handling_radio = gr.Radio(["slow_down", "loop"], label="Video Handling", value="slow_down", interactive=True)
+ fast_mode_checkbox = gr.Checkbox(label="Fast Mode (Use BiRefNet_lite)", value=False, interactive=True)
+
 
  def update_visibility(bg_type):
  if bg_type == "Color":
@@ -201,15 +223,9 @@ with gr.Blocks(theme=gr.themes.Ocean()) as demo:
 
  submit_button.click(
  fn,
- inputs=[in_video, bg_type, bg_image, bg_video, color_picker, fps_slider, video_handling_radio],
+ inputs=[in_video, bg_type, bg_image, bg_video, color_picker, fps_slider, video_handling_radio, fast_mode_checkbox],
  outputs=[stream_image, out_video],
  )
 
 if __name__ == "__main__":
- demo.launch(show_error=True)
- 
- schedule.every(10).minutes.do(clear_temp_directory)
-
- while True:
- schedule.run_pending()
- time.sleep(1)
+ demo.launch(show_error=True)