Update app.py

app.py

@@ -1,143 +1,43 @@
-import gradio as gr
-import numpy as np
-import random
-from diffusers import DiffusionPipeline
 import torch
+from diffusers import AnimateDiffPipeline, MotionAdapter, EulerDiscreteScheduler
+from diffusers.utils import export_to_gif
+from huggingface_hub import hf_hub_download
+from safetensors.torch import load_file
+import gradio as gr
 
-# Set the device based on availability
 device = "cuda" if torch.cuda.is_available() else "cpu"
-
-# Use the ByteDance/AnimateDiff-Lightning model
-model_repo_id = "ByteDance/AnimateDiff-Lightning" 
-
-# Set the torch dtype based on available hardware
-if torch.cuda.is_available():
-    torch_dtype = torch.float16
-else:
-    torch_dtype = torch.float32
-
-# Load the pipeline from the pretrained model repository
-pipe = DiffusionPipeline.from_pretrained(model_repo_id, torch_dtype=torch_dtype)
-pipe = pipe.to(device)
-
-# Maximum values for seed and image size
-MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
-
-# Define the inference function
-def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, progress=gr.Progress(track_tqdm=True)):
-
-    # Randomize seed if the checkbox is selected
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-        
-    generator = torch.Generator(device=device).manual_seed(seed)
-
-    # Generate the animation using the pipeline
-    animation = pipe(
-        prompt=prompt, 
-        negative_prompt=negative_prompt,
-        guidance_scale=guidance_scale, 
-        num_inference_steps=num_inference_steps, 
-        width=width, 
-        height=height,
-        generator=generator
-    ).images[0]  # Assuming the model generates images in the `.images` property
+dtype = torch.float16 if torch.cuda.is_available() else torch.float32
+
+step = 4  # Options: [1,2,4,8]
+repo = "ByteDance/AnimateDiff-Lightning"
+ckpt = f"animatediff_lightning_{step}step_diffusers.safetensors"
+base = "emilianJR/epiCRealism"
+
+adapter = MotionAdapter().to(device, dtype)
+adapter.load_state_dict(load_file(hf_hub_download(repo, ckpt), device=device))
+pipe = AnimateDiffPipeline.from_pretrained(base, motion_adapter=adapter, torch_dtype=dtype).to(device)
+pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config, timestep_spacing="trailing", beta_schedule="linear")
+
+def animate_image(prompt, guidance_scale, num_inference_steps):
+    output = pipe(prompt=prompt, guidance_scale=guidance_scale, num_inference_steps=num_inference_steps)
+    gif_path = "animation.gif"
+    export_to_gif(output.frames[0], gif_path)
+    return gif_path
+
+# Define the Gradio Interface
+with gr.Blocks() as demo:
+    gr.Markdown("# AnimateDiff API")
     
-    return animation, seed
-
-# Sample prompts for the UI
-examples = [
-    "A cat playing with a ball in a garden",
-    "A dancing astronaut in space",
-    "A flying dragon in the sky at sunset",
-]
-
-# Define CSS for styling
-css = """
-#col-container {
-    margin: 0 auto;
-    max-width: 640px;
-}
-"""
-
-# Build the Gradio UI
-with gr.Blocks(css=css) as demo:
+    with gr.Row():
+        prompt = gr.Textbox(label="Prompt", placeholder="A girl smiling", value="A girl smiling")
+        guidance_scale = gr.Slider(label="Guidance Scale", minimum=0.0, maximum=10.0, value=1.0, step=0.1)
+        num_inference_steps = gr.Slider(label="Steps", minimum=1, maximum=8, value=step, step=1)
     
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown(f"""
-        # AnimateDiff Lightning Model Text-to-Animation
-        """)
-        
-        with gr.Row():
-            prompt = gr.Text(
-                label="Prompt",
-                show_label=False,
-                max_lines=1,
-                placeholder="Enter your prompt",
-                container=False,
-            )
-            run_button = gr.Button("Run", scale=0)
-        
-        result = gr.Image(label="Generated Animation", show_label=False)
-
-        with gr.Accordion("Advanced Settings", open=False):
-            
-            negative_prompt = gr.Text(
-                label="Negative prompt",
-                max_lines=1,
-                placeholder="Enter a negative prompt",
-                visible=True,
-            )
-            
-            seed = gr.Slider(
-                label="Seed",
-                minimum=0,
-                maximum=MAX_SEED,
-                step=1,
-                value=0,
-            )
-            
-            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
-            
-            with gr.Row():
-                width = gr.Slider(
-                    label="Width",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,
-                )
-                height = gr.Slider(
-                    label="Height",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,
-                )
-            
-            with gr.Row():
-                guidance_scale = gr.Slider(
-                    label="Guidance scale",
-                    minimum=0.0,
-                    maximum=10.0,
-                    step=0.1,
-                    value=7.5,
-                )
-                num_inference_steps = gr.Slider(
-                    label="Number of inference steps",
-                    minimum=1,
-                    maximum=50,
-                    step=1,
-                    value=30,
-                )
-
-        # Example prompts for user selection
-        gr.Examples(
-            examples=examples,
-            inputs=[prompt]
-        )
+    gif_output = gr.Image(label="Generated Animation")
+    
+    # Button to run the pipeline
+    run_button = gr.Button("Generate Animation")
+    run_button.click(animate_image, inputs=[prompt, guidance_scale, num_inference_steps], outputs=gif_output)
 
-    # Create an API endpoint for the model
-    demo.api(fn=infer, inputs=[prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps], outputs=[result, seed])
-    demo.launch()
+# Launch the interface
+demo.launch()