Create app.py

app.py

@@ -0,0 +1,57 @@
+import torch
+from diffusers import HunyuanVideoPipeline, HunyuanVideoTransformer3DModel
+from PIL import Image
+import numpy as np
+import gradio as gr
+import os
+import gc
+
+
+model_id = "hunyuanvideo-community/HunyuanVideo"
+transformer = HunyuanVideoTransformer3DModel.from_pretrained(
+    model_id, subfolder="transformer", torch_dtype=torch.bfloat16
+    )
+
+
+pipe = HunyuanVideoPipeline.from_pretrained(model_id, transformer=transformer, torch_dtype=torch.float16)
+pipe.vae.enable_tiling()
+pipe.load_lora_weights("ovi054/ovimxVid")
+pipe.to("cuda")
+
+
+def generate(prompt, negative_prompt, width=1280, height=720, num_inference_steps=30, progress=gr.Progress(track_tqdm=True)):
+    try:
+        output = pipe(
+            prompt=prompt,
+            # negative_prompt=negative_prompt,
+            height=height,
+            width=width,
+            num_frames=1,
+            num_inference_steps=num_inference_steps,
+            # guidance_scale=5.0,
+        ).frames[0][0]
+        # image = (output * 255).astype(np.uint8)
+        # return Image.fromarray(image)
+        return output
+    finally:
+        # Always clear memory, even if an error occurs
+        torch.cuda.empty_cache()
+        gc.collect()
+
+
+iface = gr.Interface(
+    fn=generate,
+    inputs=[
+        gr.Textbox(label="Input prompt"),
+    ],
+    additional_inputs = [
+        gr.Textbox(label="Negative prompt", value = "Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards"),
+        gr.Slider(label="Width", minimum=480, maximum=1280, step=16, value=832),
+        gr.Slider(label="Height", minimum=480, maximum=1280, step=16, value=832),
+        gr.Slider(minimum=1, maximum=80, step=1, label="Inference Steps", value=30)
+    ],
+    outputs=gr.Image(label="output"),
+)
+
+
+iface.launch(share=True, debug=True)