app.py

import torch
import spaces
from diffusers import StableDiffusionPipeline, DDIMScheduler, AutoencoderKL
from transformers import AutoFeatureExtractor
from controlnet.callable_functions import process_single_image_both_ways,make_stylecode,use_stylecode

from huggingface_hub import hf_hub_download
import gradio as gr
import cv2
import os
ip_ckpt = hf_hub_download(repo_id="CiaraRowles/stylecodes", filename="stylecodes_sd15_v1.bin", repo_type="model")


device = "cuda"

example_path = os.path.join(os.path.dirname(__file__), 'test_images')
list_dir = os.listdir(example_path)
list_path = [os.path.join(example_path,img) for img in list_dir]
print("all paths ", list_path)
cv2.setNumThreads(1)

@spaces.GPU(enable_queue=True)
def generate_image(images, prompt, negative_prompt,stylecode,strength,additional_prompt, seed, progress=gr.Progress(track_tqdm=True)):

    
    full_prompt=prompt + additional_prompt

    image = images
    yield None
    
    base_size = 512

    # Calculate new width and height
   

    image = use_stylecode(ip_ckpt,image_path="",prompt=full_prompt,negative_prompt=negative_prompt,num_inference_steps=20,image=image,stylecode=stylecode,seed=seed)
    
    yield image

@spaces.GPU(enable_queue=True)
def make_stylecode_gui (images,  progress=gr.Progress(track_tqdm=True)):
    yield None
    code = make_stylecode(ip_ckpt,image_path=None,image=images)
    yield code[0]


def swap_to_gallery(images):
    return gr.update(value=images, visible=True), gr.update(visible=True), gr.update(visible=False)

def remove_back_to_files():
    return gr.update(visible=False), gr.update(visible=False), gr.update(visible=True)



    
css = '''
h1{margin-bottom: 0 !important}
'''
with gr.Blocks(css=css) as demo:
    gr.Markdown("# stylecodes demo")
    gr.Markdown("Demo for the [CiaraRowles/stylecodes model](https://huggingface.co./CiaraRowles/stylecodes)")
    with gr.Row():
        with gr.Column():
            files = gr.Image(
                label="Input image",
                type="pil"
            )
            uploaded_files = gr.Gallery(label="Your image", visible=False, columns=5, rows=1, height=125)
            with gr.Column(visible=False) as clear_button:
                remove_and_reupload = gr.ClearButton(value="Remove and upload new ones", components=files, size="sm")
            stylecode = gr.Textbox(label="stylecode")
            stylecode_gen = gr.Button("Make stylecode")
            
            
            prompt = gr.Textbox(label="Prompt",
                info="Try something like 'a photo of a man/woman/person'",
                placeholder="A photo of a [man/woman/person]...")
            
            negative_prompt = gr.Textbox(label="Negative Prompt", placeholder="low quality")
        
            with gr.Accordion(open=False, label="Advanced Options"):
                strength = gr.Slider(label="strength", value=0.9, step=0.05, minimum=0, maximum=1)
                additional_prompt = gr.Textbox(label="additional prompt",placeholder ="highly detailed ,simple background")
                seed = gr.Number(label="seed",value = -1)
            

            submit = gr.Button("Use stylecode")
        with gr.Column():
            gallery = gr.Gallery(label="Generated Images")
            examples = gr.Examples(
                inputs=files,
                examples_per_page=6,
                examples=list_path
            )
        stylecode_gen.click(fn=make_stylecode_gui,inputs=[files],outputs=stylecode)
        
        submit.click(fn=generate_image,
                    inputs=[files, prompt, negative_prompt,stylecode,strength,additional_prompt,seed],
                    outputs=gallery)
    
    gr.Markdown("note, this was trained primarily on digital painting esk generated images, don't expect it to work outside of that training data for the moment")
    gr.Markdown("This demo includes extra features to mitigate the implicit bias of the model and prevent explicit usage of it to generate content with faces of people, including third parties, that is not safe for all audiences, including naked or semi-naked people.")
    gr.Markdown("based on: https://huggingface.co./spaces/multimodalart/Ip-Adapter-FaceID")

demo.launch()