Spaces:
Running
on
Zero
Running
on
Zero
import gradio as gr | |
import torch | |
from diffusers import StableDiffusionXLPipeline, AutoencoderKL | |
from huggingface_hub import hf_hub_download | |
import lora | |
from time import sleep | |
import copy | |
sdxl_loras = [ | |
("pixel-art-xl.jpeg", "Pixel Art XL", "nerijs/pixel-art-xl", "pixel art", "pixel-art-xl.safetensors", True), | |
("papercut_SDXL.jpeg", "Papercut SDXL", "TheLastBen/Papercut_SDXL", "papercut", "papercut.safetensors", False), | |
("lego-minifig-xl.jpeg", "Lego Minifig XL", "nerijs/lego-minifig-xl", "lego minifig", "legominifig-v1.0-000003.safetensors", True), | |
("3d_style_4.jpeg", "3D Render Style", "goofyai/3d_render_style_xl", "3d style", "3d_render_style_xl.safetensors", True), | |
("LogoRedmond-LogoLoraForSDXL.jpeg","Logo.Redmond", "artificialguybr/LogoRedmond-LogoLoraForSDXL", "LogoRedAF", "LogoRedmond_LogoRedAF.safetensors", False), | |
("LineAni.Redmond.png", "LinearManga.Redmond", "artificialguybr/LineAniRedmond-LinearMangaSDXL", "LineAniAF", "LineAniRedmond-LineAniAF.safetensors", True), | |
("embroid.png","Embroidery Style","ostris/embroidery_style_lora_sdxl","","embroidered_style_v1_sdxl.safetensors",False), | |
("watercolor.png","Watercolor Style","ostris/watercolor_style_lora_sdxl","","watercolor_v1_sdxl.safetensors",False), | |
("crayon.png","Crayon Style","ostris/crayon_style_lora_sdxl","","crayons_v1_sdxl.safetensors",False), | |
("dog.png", "Cyborg Style", "goofyai/cyborg_style_xl", "cyborg style", "cyborg_style_xl-off.safetensors", True), | |
("ToyRedmond-ToyLoraForSDXL10.png","Toy.Redmond", "artificialguybr/ToyRedmond-ToyLoraForSDXL10", "FnkRedmAF", "ToyRedmond-FnkRedmAF.safetensors", True), | |
("voxel-xl-lora.png", "Voxel XL", "Fictiverse/Voxel_XL_Lora", "voxel style", "VoxelXL_v1.safetensors", True), | |
("pikachu.webp", "Pikachu XL", "TheLastBen/Pikachu_SDXL", "pikachu", "pikachu.safetensors", False), | |
("william_eggleston.webp", "William Eggleston Style", "TheLastBen/William_Eggleston_Style_SDXL", "by william eggleston", "wegg.safetensors", False), | |
("josef_koudelka.webp", "Josef Koudelka Style", "TheLastBen/Josef_Koudelka_Style_SDXL", "by josef koudelka", "koud.safetensors", False), | |
("corgi_brick.jpeg", "Lego BrickHeadz", "nerijs/lego-brickheadz-xl", "lego brickheadz", "legobrickheadz-v1.0-000004.safetensors", True) | |
] | |
saved_names = [hf_hub_download(repo_id, filename) for _, _, repo_id, _, filename, _ in sdxl_loras] | |
def update_selection(selected_state: gr.SelectData): | |
sleep(60) | |
lora_repo = sdxl_loras[selected_state.index][2] | |
instance_prompt = sdxl_loras[selected_state.index][3] | |
updated_text = f"### Selected: [{lora_repo}](https://huggingface.co/{lora_repo})" | |
return updated_text, instance_prompt, selected_state | |
vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16) | |
mutable_pipe = StableDiffusionXLPipeline.from_pretrained( | |
"stabilityai/stable-diffusion-xl-base-1.0", | |
vae=vae, | |
torch_dtype=torch.float16, | |
).to("cpu") | |
original_pipe = copy.deepcopy(mutable_pipe) | |
mutable_pipe.to("cuda") | |
last_lora = "" | |
last_merged = False | |
def run_lora(prompt, negative, weight, selected_state): | |
global last_lora, last_merged | |
pipe = mutable_pipe | |
if(not selected_state): | |
raise gr.Error("You must select a LoRA") | |
repo_name = sdxl_loras[selected_state.index][2] | |
weight_name = sdxl_loras[selected_state.index][4] | |
full_path_lora = saved_names[selected_state.index] | |
cross_attention_kwargs = None | |
if(last_lora != repo_name): | |
if(last_merged): | |
pipe = copy.deepcopy(original_pipe) | |
pipe.to("cuda") | |
else: | |
pipe.unload_lora_weights() | |
is_compatible = sdxl_loras[selected_state.index][5] | |
if(is_compatible): | |
pipe.load_lora_weights(full_path_lora) | |
cross_attention_kwargs={"scale": weight} | |
else: | |
for weights_file in [full_path_lora]: | |
if ";" in weights_file: | |
weights_file, multiplier = weights_file.split(";") | |
multiplier = float(weight) | |
else: | |
multiplier = 1.0 | |
lora_model, weights_sd = lora.create_network_from_weights( | |
multiplier, full_path_lora, pipe.vae, pipe.text_encoder, pipe.unet, for_inference=True | |
) | |
lora_model.merge_to(pipe.text_encoder, pipe.unet, weights_sd, torch.float16, "cuda") | |
last_merged = True | |
image = pipe( | |
prompt=prompt, negative_prompt=negative, num_inference_steps=20, guidance_scale=7.5, cross_attention_kwargs=cross_attention_kwargs).images[0] | |
last_lora = repo_name | |
return image | |
css = ''' | |
#prompt textarea{width: calc(100% - 160px);border-top-right-radius: 0px;border-bottom-right-radius: 0px;} | |
#run_button{position:absolute;margin-top: 38px;right: 0;margin-right: 0.8em;border-bottom-left-radius: 0px; | |
border-top-left-radius: 0px;} | |
#gallery{display:flex} | |
#gallery .grid-wrap{min-height: 100%;} | |
''' | |
with gr.Blocks(css=css) as demo: | |
title = gr.Markdown("# Lora The Explorer XL") | |
with gr.Row(): | |
gallery = gr.Gallery(value=[(a, b) for a, b, _, _, _, _ in sdxl_loras], | |
label="LoRA Gallery", | |
allow_preview=False, | |
columns=3, | |
elem_id="gallery" | |
) | |
with gr.Column(): | |
prompt_title = gr.Markdown(value="### Click on a LoRA in the gallery to select it", visible=True) | |
with gr.Row(): | |
prompt = gr.Textbox(label="Prompt", elem_id="prompt") | |
button = gr.Button("Run", elem_id="run_button") | |
result = gr.Image(interactive=False, label="result") | |
with gr.Accordion("Advanced options", open=False): | |
negative = gr.Textbox(label="Negative Prompt") | |
weight = gr.Slider(0, 1, value=1, step=0.1, label="LoRA weight") | |
with gr.Column(): | |
gr.Markdown("Use it with:") | |
with gr.Row(): | |
with gr.Accordion("🧨 diffusers", open=False): | |
gr.Markdown("") | |
with gr.Accordion("ComfyUI", open=False): | |
gr.Markdown("") | |
with gr.Accordion("Invoke AI", open=False): | |
gr.Markdown("") | |
with gr.Accordion("SD.Next (AUTO1111 fork)", open=False): | |
gr.Markdown("") | |
selected_state = gr.State() | |
gallery.select(update_selection, outputs=[prompt_title, prompt, selected_state], queue=False, show_progress=False) | |
button.click(fn=run_lora, inputs=[prompt, negative, weight, selected_state], outputs=result) | |
demo.launch() |