import spaces import gradio as gr import shutil import os import subprocess import sys # Run the .bat file before launching the app try: import PromptTrack except ImportError: print("PromptTrack not found. Installing...") subprocess.run([sys.executable, "-m", "pip", "install", "--index-url", "https://test.pypi.org/simple/", "--extra-index-url", "https://pypi.org/simple/", "PromptTrack"], check=True) subprocess.run([sys.executable, "-m", "pip", "install", "--no-deps", "bytetracker"], check=True) import PromptTrack # Retry import after installation from PromptTrack import PromptTracker tracker = PromptTracker() #@spaces.GPU(duration=300) def process_video(video_path, prompt): import torch print(f"Is CUDA available: {torch.cuda.is_available()}") # True print(f"CUDA device: {torch.cuda.get_device_name(torch.cuda.current_device())}") # Tesla T4 detection_threshold=0.3 track_thresh=0.4 match_thresh=1 max_time_lost=float("inf") nbr_frames_fixing=800 output_video = video_path.split('mp4')[0]+"_with_id.mp4" # Placeholder for processed video output_file = video_path.split('mp4')[0]+"_mot_.json" # Tracking result output_file_2 = video_path.split('mp4')[0]+"_object_detection.json" # detection results video_file = video_path tracker.detect_objects(video_file, prompt=prompt, nms_threshold=0.8, detection_threshold=detection_threshold, detector="OWL-VITV2") tracker.process_mot(video_file, fixed_parc=True, track_thresh=track_thresh, match_thresh=match_thresh, frame_rate=25, max_time_lost=max_time_lost, nbr_frames_fixing=nbr_frames_fixing) tracker.read_video_with_mot(video_file, fps=25) """output_video = "output.mp4" # Placeholder for processed video output_file = "output.txt" # Placeholder for generated file # Copy the input video to simulate processing shutil.copy(video_path.name, output_video) # Create an output text file with the prompt content with open(output_file, "w") as f: f.write(f"User Prompt: {prompt}\n") """ return output_video, output_file # Define Gradio interface iface = gr.Interface( fn=process_video, inputs=[gr.File(label="Upload Video"), gr.Textbox(placeholder="Enter your prompt")], outputs=[gr.Video(), gr.File(label="Generated File")], title="Video Processing App", description="Upload a video and enter a prompt. The app will return the processed video and a generated file." ) # Launch the app if __name__ == "__main__": iface.launch(share=True)