Update app.py

app.py

@@ -6,7 +6,7 @@ from inference.flovd_demo import generate_video
 from huggingface_hub import snapshot_download
 
 # -----------------------------------
-# Step 1: Setup Hugging Face Cache
+# Step 1: Download model checkpoints
 # -----------------------------------
 hf_token = os.getenv("HF_TOKEN", None)
 
@@ -18,16 +18,6 @@ snapshot_download(
     token=hf_token,
 )
 
-# Optional: print ckpt directory structure
-print("📁 CKPT Directory Contents:")
-for root, dirs, files in os.walk("./ckpt"):
-    level = root.replace("./ckpt", "").count(os.sep)
-    indent = " " * 4 * level
-    print(f"{indent}{os.path.basename(root)}/")
-    subindent = " " * 4 * (level + 1)
-    for f in files:
-        print(f"{subindent}{f}")
-
 # -----------------------------------
 # Step 2: Define inference function
 # -----------------------------------
@@ -39,6 +29,11 @@ OUTPUT_PATH = "./results/"
 POSE_TYPE = "re10k"
 CONTROLNET_GUIDANCE_END = 0.4
 SPEED = 1.0
+NUM_FRAMES = 81
+FPS = 16
+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"
@@ -51,18 +46,27 @@ def run_flovd(prompt, image, cam_pose_name):
         image_path=image_path,
         cam_pose_name=cam_pose_name,
         output_path=OUTPUT_PATH,
-        pose_type=POSE_TYPE,
         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,
     )
 
-    return "./results/generated_videos/your_video.mp4"
+    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):
+        return video_path
+    else:
+        return "Video generation failed or file not found."
 
 # -----------------------------------
-# Step 3: Gradio UI
+# Step 3: Launch Gradio UI
 # -----------------------------------
 
 iface = gr.Interface(
@@ -70,10 +74,11 @@ iface = gr.Interface(
     inputs=[
         gr.Textbox(label="Prompt"),
         gr.Image(type="pil", label="Input Image"),
-        gr.Textbox(label="Camera Pose File Name", value="example.txt"),
+        gr.Textbox(label="Camera Pose File Name (e.g., 1593596b99e2dde9.txt)"),
     ],
     outputs=gr.Video(label="Generated Video"),
-    title="FloVD - Camera Motion Guided Video Generation"
+    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."
 )
 
 iface.launch(server_name="0.0.0.0", server_port=7860)