# PyTorch 2.8 (temporary hack) import os os.system('pip install --upgrade --pre --extra-index-url https://download.pytorch.org/whl/nightly/cu126 "torch<2.9"') from huggingface_hub import HfApi, upload_file import uuid import subprocess import tempfile import logging import shutil from datetime import datetime import spaces import torch from diffusers.pipelines.wan.pipeline_wan_i2v import WanImageToVideoPipeline from diffusers.models.transformers.transformer_wan import WanTransformer3DModel from diffusers.utils.export_utils import export_to_video import gradio as gr import numpy as np from PIL import Image import random import gc from optimization import optimize_pipeline_ from huggingface_hub import hf_hub_download MODEL_ID = "Wan-AI/Wan2.2-I2V-A14B-Diffusers" LORA_REPO_ID = "rahul7star/wan2.2Lora" LORA_SETS = { "NF": { "high_noise": {"file": "NSFW-22-H-e8.safetensors", "adapter_name": "nf_high"}, "low_noise": {"file": "NSFW-22-L-e8.safetensors", "adapter_name": "nf_low"} }, "BP": { "high_noise": {"file": "Wan2.2_BP-v1-HighNoise-I2V_T2V.safetensors", "adapter_name": "bp_high"}, "low_noise": {"file": "Wan2.2_BP-v1-LowNoise-I2V_T2V.safetensors", "adapter_name": "bp_low"} }, "Py-v1": { "high_noise": {"file": "wan2.2_i2v_highnoise_pov_missionary_v1.0.safetensors", "adapter_name": "py_high"}, "low_noise": {"file": "wan2.2_i2v_lownoise_pov_missionary_v1.0.safetensors", "adapter_name": "py_low"} } } LANDSCAPE_WIDTH = 832 LANDSCAPE_HEIGHT = 576 MAX_SEED = np.iinfo(np.int32).max FIXED_FPS = 16 MIN_FRAMES_MODEL = 8 MAX_FRAMES_MODEL = 81 MIN_DURATION = round(MIN_FRAMES_MODEL / FIXED_FPS, 1) MAX_DURATION = round(MAX_FRAMES_MODEL / FIXED_FPS, 1) # ---------------- Pipeline ----------------- pipe = WanImageToVideoPipeline.from_pretrained( MODEL_ID, transformer=WanTransformer3DModel.from_pretrained( 'cbensimon/Wan2.2-I2V-A14B-bf16-Diffusers', subfolder='transformer', torch_dtype=torch.bfloat16, device_map='cuda', ), transformer_2=WanTransformer3DModel.from_pretrained( 'cbensimon/Wan2.2-I2V-A14B-bf16-Diffusers', subfolder='transformer_2', torch_dtype=torch.bfloat16, device_map='cuda', ), torch_dtype=torch.bfloat16, ).to('cuda') # Optimize once for AoT optimize_pipeline_( pipe, image=Image.new('RGB', (LANDSCAPE_WIDTH, LANDSCAPE_HEIGHT)), prompt='prompt', height=LANDSCAPE_HEIGHT, width=LANDSCAPE_WIDTH, num_frames=MAX_FRAMES_MODEL, ) # ---------------- Load LoRA Weights ----------------- for name, lora_set in LORA_SETS.items(): print(f"--- LoRA 集合: {name} ---") high_noise_config = lora_set["high_noise"] print(f"High Noise: {high_noise_config['file']}...") pipe.load_lora_weights( LORA_REPO_ID, weight_name=high_noise_config['file'], adapter_name=high_noise_config['adapter_name'] ) print("High Noise LoRA 加载完成。") low_noise_config = lora_set["low_noise"] print(f"Low Noise: {low_noise_config['file']}...") pipe.load_lora_weights( LORA_REPO_ID, weight_name=low_noise_config['file'], adapter_name=low_noise_config['adapter_name'] ) print("Low Noise LoRA 加载完成。") # Fuse once globally try: pipe.fuse_lora() print("✅ 全局 Fuse LoRA 成功") except Exception as e: print(f"⚠️ Fuse LoRA 失败: {e}") # Clean GPU for i in range(3): gc.collect() torch.cuda.synchronize() torch.cuda.empty_cache() # ---------------- Defaults ----------------- default_prompt_i2v = "make this image come alive, cinematic motion, smooth animation" default_negative_prompt = ( "色调艳丽, 过曝, 静态, 细节模糊不清, 字幕, 风格, 作品, 画作, 画面, 静止, " "整体发灰, 最差质量, 低质量, JPEG压缩残留, 丑陋的, 残缺的, 多余的手指, " "画得不好的手部, 画得不好的脸部, 畸形的, 毁容的, 形态畸形的肢体, 手指融合, " "静止不动的画面, 杂乱的背景, 三条腿, 背景人很多, 倒着走" ) # ---------------- Utils ----------------- def resize_image(image: Image.Image) -> Image.Image: if image.height > image.width: transposed = image.transpose(Image.Transpose.ROTATE_90) resized = resize_image_landscape(transposed) return resized.transpose(Image.Transpose.ROTATE_270) return resize_image_landscape(image) def resize_image_landscape(image: Image.Image) -> Image.Image: target_aspect = LANDSCAPE_WIDTH / LANDSCAPE_HEIGHT width, height = image.size in_aspect = width / height if in_aspect > target_aspect: new_width = round(height * target_aspect) left = (width - new_width) // 2 image = image.crop((left, 0, left + new_width, height)) else: new_height = round(width / target_aspect) top = (height - new_height) // 2 image = image.crop((0, top, width, top + new_height)) return image.resize((LANDSCAPE_WIDTH, LANDSCAPE_HEIGHT), Image.LANCZOS) def get_duration( input_image, prompt, steps, negative_prompt, duration_seconds, guidance_scale, guidance_scale_2, seed, randomize_seed, selected_loras, progress, ): return int(steps) * 15 # ---------------- LoRA Switcher ----------------- class LoraSwitcher: def __init__(self, selected_lora_names, switch_step): self.switched = False self.high_noise_adapters = [] self.low_noise_adapters = [] self.switch_step = switch_step if selected_lora_names: for name in selected_lora_names: if name in LORA_SETS: self.high_noise_adapters.append(LORA_SETS[name]["high_noise"]["adapter_name"]) self.low_noise_adapters.append(LORA_SETS[name]["low_noise"]["adapter_name"]) def __call__(self, pipe, step_index, timestep, callback_kwargs): if step_index == 0: self.switched = False if self.high_noise_adapters: print(f"激活 High Noise LoRA: {self.high_noise_adapters}") pipe.set_adapters(self.high_noise_adapters, adapter_weights=[1.0]*len(self.high_noise_adapters)) try: pipe.fuse_lora() print("Fuse High Noise LoRA ✅") except Exception as e: print(f"Fuse High Noise LoRA 失败: {e}") elif pipe.get_active_adapters(): active = pipe.get_active_adapters() print(f"禁用残留的 LoRA: {active}") pipe.set_adapters(active, adapter_weights=[0.0]*len(active)) if self.low_noise_adapters and step_index >= self.switch_step and not self.switched: print(f"在第 {step_index} 步切换到 Low Noise LoRA: {self.low_noise_adapters}") pipe.set_adapters(self.low_noise_adapters, adapter_weights=[1.0]*len(self.low_noise_adapters)) try: pipe.fuse_lora() print("Fuse Low Noise LoRA ✅") except Exception as e: print(f"Fuse Low Noise LoRA 失败: {e}") self.switched = True return callback_kwargs # ---------------- Main Generation ----------------- @spaces.GPU(duration=get_duration) def generate_video( input_image, prompt, steps=4, negative_prompt=default_negative_prompt, duration_seconds=MAX_DURATION, guidance_scale=1, guidance_scale_2=1, seed=42, randomize_seed=False, selected_loras=[], progress=gr.Progress(track_tqdm=True), ): if input_image is None: raise gr.Error("Please upload an input image.") print("Prompt is:", prompt) # Reset fused LoRA before new run try: pipe.unfuse_lora() print("🔄 Reset unfuse_lora before generation") except Exception: pass num_frames = np.clip(int(round(duration_seconds * FIXED_FPS)), MIN_FRAMES_MODEL, MAX_FRAMES_MODEL) current_seed = random.randint(0, MAX_SEED) if randomize_seed else int(seed) resized_image = resize_image(input_image) num_inference_steps = int(steps) switch_step = num_inference_steps // 2 lora_switcher_callback = LoraSwitcher(selected_loras, switch_step) output_frames_list = pipe( image=resized_image, prompt=prompt, negative_prompt=negative_prompt, height=resized_image.height, width=resized_image.width, num_frames=num_frames, guidance_scale=float(guidance_scale), guidance_scale_2=float(guidance_scale_2), num_inference_steps=num_inference_steps, generator=torch.Generator(device="cuda").manual_seed(current_seed), callback_on_step_end=lora_switcher_callback, ).frames[0] with tempfile.NamedTemporaryFile(suffix=".mp4", delete=False) as tmpfile: video_path = tmpfile.name export_to_video(output_frames_list, video_path, fps=FIXED_FPS) return video_path, current_seed # ---------------- UI ----------------- with gr.Blocks() as demo: gr.Markdown("# Fast 4 steps Wan 2.2 I2V (14B) with Lightning LoRA") gr.Markdown("Run Wan 2.2 in just 4-8 steps, with Lightning LoRA, fp8 quantization & AoT compilation - compatible with 🧨 diffusers and ZeroGPU⚡️") with gr.Row(): with gr.Column(): input_image_component = gr.Image(type="pil", label="Input Image (auto-resized)", interactive=True) prompt_input = gr.Textbox(label="Prompt", value=default_prompt_i2v) duration_seconds_input = gr.Slider(minimum=MIN_DURATION, maximum=MAX_DURATION, step=0.1, value=3.5, label="Duration (seconds)") lora_selection_checkbox = gr.CheckboxGroup(choices=list(LORA_SETS.keys()), label="选择要应用的 LoRA (可多选)") with gr.Accordion("Advanced Settings", open=False): negative_prompt_input = gr.Textbox(label="Negative Prompt", value=default_negative_prompt, lines=3) seed_input = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=42) randomize_seed_checkbox = gr.Checkbox(label="Randomize seed", value=True) steps_slider = gr.Slider(minimum=1, maximum=30, step=1, value=6, label="Inference Steps") guidance_scale_input = gr.Slider(minimum=0.0, maximum=10.0, step=0.5, value=1, label="Guidance Scale - high noise") guidance_scale_2_input = gr.Slider(minimum=0.0, maximum=10.0, step=0.5, value=1, label="Guidance Scale 2 - low noise") generate_button = gr.Button("Generate Video", variant="primary") with gr.Column(): video_output = gr.Video(label="Generated Video", autoplay=True) ui_inputs = [ input_image_component, prompt_input, steps_slider, negative_prompt_input, duration_seconds_input, guidance_scale_input, guidance_scale_2_input, seed_input, randomize_seed_checkbox, lora_selection_checkbox ] generate_button.click(fn=generate_video, inputs=ui_inputs, outputs=[video_output, seed_input]) if __name__ == "__main__": demo.queue().launch()