Update app.py

app.py

@@ -22,6 +22,9 @@ image_encoder = CLIPVisionModel.from_pretrained(
 
 print("###### Loading VAE encoder ######")
 vae = AutoencoderKLWan.from_pretrained(model_id, subfolder="vae", torch_dtype=torch.float32)
+
+print("Loading Pipeline...")
+
 pipe = WanImageToVideoPipeline.from_pretrained(
     model_id, vae=vae, image_encoder=image_encoder, torch_dtype=torch.bfloat16
 )
@@ -44,12 +47,21 @@ except:
     print("Model CPU Offload failed")
 
 
-
+try:
+    print("Enabling Attention Slicing ")
+    pipe.enable_attention_slicing()
+    print("Attention Slicing Enabled")
+except Exception as e:
+    print("Attention Slicing Failed")
 
 
 # Loading function for Image
 from diffusers.utils import load_image
 
+# ================================
+# Image Preparation Function
+# ================================
+
 def prepare_vertical_image(pipe, image_path, base_width=384, base_height=672):
     """
     Loads and resizes an image for Wan I2V vertical video generation.
@@ -84,20 +96,70 @@ def prepare_vertical_image(pipe, image_path, base_width=384, base_height=672):
 
 
 
-# how to use the Image loading
-image, width, height = prepare_vertical_image(
-    pipe,
-    "input.jpg",
-    base_width=384,
-    base_height=672
-)
 
+@spaces.GPU(size="xlarge", duration=get_duration)
+def generate_video(input_image, prompt, negative_prompt):
+
+    if input_image is None:
+        return None
+
+    image = input_image
+
+    # Prepare 9:16 vertical reduced resolution
+    image, width, height = prepare_vertical_image(pipe, image)
+
+    print(f"Generating 10 sec vertical video at {width}x{height}")
+
+    # 10 seconds at 16 FPS = 160 frames
+    video_frames = pipe(
+        image=image,
+        prompt=prompt,
+        negative_prompt=negative_prompt,
+        height=height,
+        width=width,
+        num_frames=160,
+        guidance_scale=4.5,
+        num_inference_steps=25
+    ).frames[0]
+
+    output_path = "vertical_output.mp4"
+    export_to_video(video_frames, output_path, fps=16)
+
+    return output_path
+
+
+# Gradio UI
+# ================================
+
+with gr.Blocks(title="Wan 14B Vertical I2V") as demo:
+
+    gr.Markdown("## 🎬 Wan 14B Image-to-Video Generator")
+    gr.Markdown("Generate 10-second Vertical (9:16) AI Videos")
+
+    with gr.Row():
+        input_image = gr.Image(type="pil", label="Upload Image")
+    
+    prompt = gr.Textbox(
+        label="Prompt",
+        placeholder="Describe motion, camera movement, cinematic effect..."
+    )
+
+    negative_prompt = gr.Textbox(
+        label="Negative Prompt",
+        value="blurry, low quality, distorted, static",
+    )
+
+    generate_btn = gr.Button("Generate 10 Second Video")
+
+    output_video = gr.Video(label="Generated Video")
+
+    generate_btn.click(
+        generate_video,
+        inputs=[input_image, prompt, negative_prompt],
+        outputs=output_video
+    )
 
+demo.launch(server_name="0.0.0.0", server_port=7860)
 
 
-@spaces.GPU(duration=60)
-def greet(name):
-    return "Hello " + name + "!!"
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()