app.py

import gradio as gr
import jax
import numpy as np
import jax.numpy as jnp
from flax.jax_utils import replicate
from flax.training.common_utils import shard
from PIL import Image
from diffusers import FlaxStableDiffusionControlNetPipeline, FlaxControlNetModel
import cv2
import os





def image_grid(imgs, rows, cols):
    w, h = imgs[0].size
    grid = Image.new("RGB", size=(cols * w, rows * h))
    for i, img in enumerate(imgs):
        grid.paste(img, box=(i % cols * w, i // cols * h))
    return grid


def create_key(seed=0):
    return jax.random.PRNGKey(seed)




def canny_filter(image):
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    
    blurred_image = cv2.GaussianBlur(gray_image, (5, 5), 0)
    edges_image = cv2.Canny(blurred_image, 50, 150)
    return edges_image

# load control net and stable diffusion v1-5
controlnet, controlnet_params = FlaxControlNetModel.from_pretrained(
    "jax-diffusers-event/canny-coyo1m", dtype=jnp.float32
)
pipe, params = FlaxStableDiffusionControlNetPipeline.from_pretrained(
    "runwayml/stable-diffusion-v1-5", controlnet=controlnet, from_pt=True, dtype=jnp.float32
)

def infer(prompts, negative_prompts, image):
    
    
    params["controlnet"] = controlnet_params
    
    num_samples = jax.device_count()
    rng = create_key(0)
    rng = jax.random.split(rng, jax.device_count())
    canny_image = canny_filter(image)
    
    prompt_ids = pipe.prepare_text_inputs([prompts] * num_samples)
    negative_prompt_ids = pipe.prepare_text_inputs([negative_prompts] * num_samples)
    processed_image = pipe.prepare_image_inputs([canny_image] * num_samples)
    
    p_params = replicate(params)
    prompt_ids = shard(prompt_ids)
    negative_prompt_ids = shard(negative_prompt_ids)
    processed_image = shard(processed_image)
    
    output = pipe(
        prompt_ids=prompt_ids,
        image=processed_image,
        params=p_params,
        prng_seed=rng,
        num_inference_steps=50,
        neg_prompt_ids=negative_prompt_ids,
        jit=True,
    ).images
    
    output_images = pipe.numpy_to_pil(np.asarray(output.reshape((num_samples,) + output.shape[-3:])))
    output_images = image_grid(output_images, num_samples // 4, 4)
    return output_images

gr.Interface(infer, inputs=["text", "text", "image"], outputs="gallery").launch()