import spaces import argparse import os import time from os import path from PIL import ImageOps cache_path = path.join(path.dirname(path.abspath(__file__)), "models") os.environ["TRANSFORMERS_CACHE"] = cache_path os.environ["HF_HUB_CACHE"] = cache_path os.environ["HF_HOME"] = cache_path import gradio as gr import torch from diffusers import StableDiffusionControlNetPipeline, ControlNetModel from scheduling_tcd import TCDScheduler torch.backends.cuda.matmul.allow_tf32 = True js_func = """ function refresh() { const url = new URL(window.location); if (url.searchParams.get('__theme') !== 'dark') { url.searchParams.set('__theme', 'dark'); window.location.href = url.href; } } """ class timer: def __init__(self, method_name="timed process"): self.method = method_name def __enter__(self): self.start = time.time() print(f"{self.method} starts") def __exit__(self, exc_type, exc_val, exc_tb): end = time.time() print(f"{self.method} took {str(round(end - self.start, 2))}s") if not path.exists(cache_path): os.makedirs(cache_path, exist_ok=True) controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-scribble", torch_dtype=torch.float16, use_safetensors=True) pipe = StableDiffusionControlNetPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", controlnet=controlnet, torch_dtype=torch.float16, variant="fp16") pipe.load_lora_weights("ByteDance/Hyper-SD", weight_name="Hyper-SD15-1step-lora.safetensors", adapter_name="default") pipe.to("cuda") pipe.scheduler = TCDScheduler.from_config(pipe.scheduler.config, timestep_spacing ="trailing") with gr.Blocks(js=js_func) as demo: with gr.Column(): with gr.Row(): with gr.Column(): # scribble = gr.Image(source="canvas", tool="color-sketch", shape=(512, 512), height=768, width=768, type="pil") scribble = gr.ImageEditor(type="pil", image_mode="L", crop_size=(512, 512), sources=(), brush=gr.Brush(color_mode="fixed", colors=["#FFFFFF"])) # scribble_out = gr.Image(height=384, width=384) num_images = gr.Slider(label="Number of Images", minimum=1, maximum=8, step=1, value=4, interactive=True) steps = gr.Slider(label="Inference Steps", minimum=1, maximum=8, step=1, value=1, interactive=True) prompt = gr.Text(label="Prompt", value="a photo of a cat", interactive=True) eta = gr.Number(label="Eta (Corresponds to parameter eta (η) in the DDIM paper, i.e. 0.0 eqauls DDIM, 1.0 equals LCM)", value=1., interactive=True) controlnet_scale = gr.Number(label="ControlNet Conditioning Scale", value=1.0, interactive=True) seed = gr.Number(label="Seed", value=3413, interactive=True) btn = gr.Button(value="run") with gr.Column(): output = gr.Gallery(height=768, format="png") # output = gr.Image() @spaces.GPU def process_image(steps, prompt, controlnet_scale, eta, seed, scribble, num_images): global pipe if scribble: with torch.inference_mode(), torch.autocast("cuda", dtype=torch.float16), timer("inference"): result = pipe( prompt=[prompt]*num_images, image=[ImageOps.invert(scribble['composite'])]*num_images, # image=[scribble['composite']]*num_images, generator=torch.Generator().manual_seed(int(seed)), num_inference_steps=steps, guidance_scale=0., eta=eta, controlnet_conditioning_scale=float(controlnet_scale), ).images # result[0].save("test.jpg") # print(result[0]) return result else: return None reactive_controls = [steps, prompt, controlnet_scale, eta, seed, scribble, num_images] for control in reactive_controls: if reactive_controls[-2] is not None: control.change(fn=process_image, inputs=reactive_controls, outputs=[output, ]) btn.click(process_image, inputs=reactive_controls, outputs=[output, ]) if __name__ == "__main__": demo.launch()