Update app.py

app.py

@@ -17,9 +17,9 @@ snapshot_download(
     allow_patterns="ckpt/**",
     token=hf_token,
 )
-print("ckpt dir files:", os.listdir("./ckpt"))
 
-print("\n📂 Contents of ./ckpt directory:")
+print("✅ Download complete.")
+print("📂 Contents of ./ckpt directory:")
 for root, dirs, files in os.walk("./ckpt"):
     level = root.replace("./ckpt", "").count(os.sep)
     indent = " " * 4 * level
@@ -28,7 +28,6 @@ for root, dirs, files in os.walk("./ckpt"):
     for f in files:
         print(f"{subindent}{f}")
 
-
 # -----------------------------------
 # Step 2: Define inference function
 # -----------------------------------
@@ -47,35 +46,48 @@ INFER_STEPS = 50
 os.makedirs(os.path.join(OUTPUT_PATH, "generated_videos"), exist_ok=True)
 
 def run_flovd(prompt, image, cam_pose_name):
-    image_path = "./temp_input.png"
-    image.save(image_path)
-    print("image : ", image)
+    try:
+        # Save input image
+        image_path = "./temp_input.png"
+        image.save(image_path)
+        print(f"📸 Image saved at {image_path}")
+
+        # Run video generation
+        generate_video(
+            prompt=prompt,
+            fvsm_path=FVSM_PATH,
+            omsm_path=OMSM_PATH,
+            image_path=image_path,
+            cam_pose_name=cam_pose_name,
+            output_path=OUTPUT_PATH,
+            controlnet_guidance_end=CONTROLNET_GUIDANCE_END,
+            pose_type=POSE_TYPE,
+            speed=SPEED,
+            use_flow_integration=True,
+            depth_ckpt_path=DEPTH_CKPT_PATH,
+            dtype=torch.float16,
+            num_frames=NUM_FRAMES,
+            fps=FPS,
+            num_inference_steps=INFER_STEPS,
+        )
+
+        # Construct video path
+        prompt_short = prompt[:30].strip().replace(" ", "_").replace(".", "").replace(",", "")
+        video_filename = f"{prompt_short}_{cam_pose_name}.mp4"
+        video_path = os.path.join(OUTPUT_PATH, "generated_videos", video_filename)
 
-    generate_video(
-        prompt=prompt,
-        fvsm_path=FVSM_PATH,
-        omsm_path=OMSM_PATH,
-        image_path=image_path,
-        cam_pose_name=cam_pose_name,
-        output_path=OUTPUT_PATH,
-        controlnet_guidance_end=CONTROLNET_GUIDANCE_END,
-        pose_type=POSE_TYPE,
-        speed=SPEED,
-        use_flow_integration=True,
-        depth_ckpt_path=DEPTH_CKPT_PATH,
-        dtype=torch.float16,
-        num_frames=NUM_FRAMES,
-        fps=FPS,
-        num_inference_steps=INFER_STEPS,
-    )
+        print(f"📍 Looking for generated video at: {video_path}")
 
-    prompt_short = prompt[:30].strip().replace(" ", "_").replace(".", "").replace(",", "")
-    video_path = os.path.join(OUTPUT_PATH, "generated_videos", f"{prompt_short}_{cam_pose_name}.mp4")
+        if os.path.exists(video_path):
+            print("✅ Video generation successful.")
+            return video_path
+        else:
+            print("❌ Video generation failed or file not found.")
+            return f"Video generation failed. Expected output file not found at: {video_path}"
 
-    if os.path.exists(video_path):
-        return video_path
-    else:
-        return "Video generation failed or file not found."
+    except Exception as e:
+        print(f"🔥 Exception occurred: {str(e)}")
+        return f"An error occurred during generation: {str(e)}"
 
 # -----------------------------------
 # Step 3: Launch Gradio UI
@@ -88,9 +100,9 @@ iface = gr.Interface(
         gr.Image(type="pil", label="Input Image"),
         gr.Textbox(label="Camera Pose File Name (e.g., 1593596b99e2dde9.txt)"),
     ],
-    outputs=gr.Video(label="Generated Video"),
+    outputs=gr.Video(label="Generated Video or Error Message"),
     title="🎥 FloVD + CogVideoX - Camera Motion Guided Video Generation",
-    description="Upload an image, enter a descriptive prompt, and provide a camera pose file name to generate video with dynamic camera motion."
+    description="Upload an image, enter a prompt, and provide a camera pose filename to generate a video."
 )
 
 iface.launch(server_name="0.0.0.0", server_port=7860)