app.py

import spaces
import logging
import math
import gradio as gr
from PIL import Image
from src.tryon_pipeline import StableDiffusionXLInpaintPipeline as TryonPipeline
from src.unet_hacked_garmnet import UNet2DConditionModel as UNet2DConditionModel_ref
from src.unet_hacked_tryon import UNet2DConditionModel
from src.enhanced_garment_net import EnhancedGarmentNetWithTimestep
from transformers import (
    CLIPImageProcessor,
    CLIPVisionModelWithProjection,
    CLIPTextModel,
    CLIPTextModelWithProjection,
)
from diffusers import DDPMScheduler,AutoencoderKL
from typing import List

import torch
import os
from transformers import AutoTokenizer
import numpy as np
from utils_mask import get_mask_location
from torchvision import transforms
import apply_net
from preprocess.humanparsing.run_parsing import Parsing
from preprocess.openpose.run_openpose import OpenPose
from detectron2.data.detection_utils import convert_PIL_to_numpy,_apply_exif_orientation
from torchvision.transforms.functional import to_pil_image
from src.background_processor import BackgroundProcessor

def pil_to_binary_mask(pil_image, threshold=0):
    np_image = np.array(pil_image)
    grayscale_image = Image.fromarray(np_image).convert("L")
    binary_mask = np.array(grayscale_image) > threshold
    mask = np.zeros(binary_mask.shape, dtype=np.uint8)
    for i in range(binary_mask.shape[0]):
        for j in range(binary_mask.shape[1]):
            if binary_mask[i,j] == True :
                mask[i,j] = 1
    mask = (mask*255).astype(np.uint8)
    output_mask = Image.fromarray(mask)
    return output_mask


base_path = 'yisol/IDM-VTON'
example_path = os.path.join(os.path.dirname(__file__), 'example')

unet = UNet2DConditionModel.from_pretrained(
    base_path,
    subfolder="unet",
    torch_dtype=torch.float16,
)
unet.requires_grad_(False)

# This is suggestion from Claude for enhanced garment net
#enhancedGarmentNet = EnhancedGarmentNetWithTimestep()
#enhancedGarmentNet.to(dtype=torch.float16)

tokenizer_one = AutoTokenizer.from_pretrained(
    base_path,
    subfolder="tokenizer",
    revision=None,
    use_fast=False,
)
tokenizer_two = AutoTokenizer.from_pretrained(
    base_path,
    subfolder="tokenizer_2",
    revision=None,
    use_fast=False,
)
noise_scheduler = DDPMScheduler.from_pretrained(base_path, subfolder="scheduler")

text_encoder_one = CLIPTextModel.from_pretrained(
    base_path,
    subfolder="text_encoder",
    torch_dtype=torch.float16,
)
text_encoder_two = CLIPTextModelWithProjection.from_pretrained(
    base_path,
    subfolder="text_encoder_2",
    torch_dtype=torch.float16,
)
image_encoder = CLIPVisionModelWithProjection.from_pretrained(
    base_path,
    subfolder="image_encoder",
    torch_dtype=torch.float16,
    )
vae = AutoencoderKL.from_pretrained(base_path,
                                    subfolder="vae",
                                    torch_dtype=torch.float16,
)

# "stabilityai/stable-diffusion-xl-base-1.0",
UNet_Encoder = UNet2DConditionModel_ref.from_pretrained(
    base_path,
    subfolder="unet_encoder",
    torch_dtype=torch.float16,
)


parsing_model = Parsing(0)
openpose_model = OpenPose(0)

UNet_Encoder.requires_grad_(False)
image_encoder.requires_grad_(False)
vae.requires_grad_(False)
unet.requires_grad_(False)
text_encoder_one.requires_grad_(False)
text_encoder_two.requires_grad_(False)
tensor_transfrom = transforms.Compose(
            [
                transforms.ToTensor(),
                transforms.Normalize([0.5], [0.5]),
            ]
    )

pipe = TryonPipeline.from_pretrained(
        base_path,
        unet=unet,
        vae=vae,
        feature_extractor= CLIPImageProcessor(),
        text_encoder = text_encoder_one,
        text_encoder_2 = text_encoder_two,
        tokenizer = tokenizer_one,
        tokenizer_2 = tokenizer_two,
        scheduler = noise_scheduler,
        image_encoder=image_encoder,
        torch_dtype=torch.float16,
)
pipe.unet_encoder = UNet_Encoder
# pipe.garment_net = enhancedGarmentNet

# Standard size of shein images
#WIDTH = int(4160/5)
#HEIGHT = int(6240/5)
# Standard size on which model is trained
WIDTH = int(768)
HEIGHT = int(1024)
POSE_WIDTH = int(WIDTH/2)  # int(WIDTH/2)
POSE_HEIGHT = int(HEIGHT/2)  #int(HEIGHT/2)
ARM_WIDTH = "dc" # "hd" # hd -> full sleeve, dc for half sleeve 
CATEGORY = "upper_body" # "lower_body"


def is_cropping_required(width, height):
    # If aspect ratio is 1.33, which is same as standard 3x4 ( 768x1024 ), then no need to crop, else crop
    aspect_ratio = round(height/width, 2)
    if aspect_ratio == 1.33:
        return False
    return True


@spaces.GPU
def start_tryon(human_img_dict,garm_img,garment_des, background_img, is_checked,is_checked_crop,denoise_steps,seed):
    logging.info("Starting try on")
    print(f"Input: {human_img_dict}")
    #device = "cuda"
    device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
    
    openpose_model.preprocessor.body_estimation.model.to(device)
    pipe.to(device)
    pipe.unet_encoder.to(device)
    # pipe.garment_net.to(device)
    
    # human_img_orig = human_img_dict["background"].convert("RGB")   # ImageEditor
    human_img_orig = human_img_dict.convert("RGB")     # Image
    

    """
    # Derive HEIGHT & WIDTH such that width is not more than 1000. This will cater to both Shein images (4160x6240) of 2:3 AR and model standard images ( 768x1024 ) of 3:4 AR
    WIDTH, HEIGHT = human_img_orig.size
    division_factor = math.ceil(WIDTH/1000)
    WIDTH = int(WIDTH/division_factor)
    HEIGHT = int(HEIGHT/division_factor)
    POSE_WIDTH = int(WIDTH/2)
    POSE_HEIGHT = int(HEIGHT/2)
    """
    # is_checked_crop as True if original AR is not same as 2x3 as expected by model
    w, h = human_img_orig.size
    is_checked_crop = is_cropping_required(w, h)

    garm_img= garm_img.convert("RGB").resize((WIDTH,HEIGHT))
    if is_checked_crop:
        # This will crop the image to make it Aspect Ratio of 3 x 4. And then at the end revert it back to original dimentions
        width, height = human_img_orig.size
        target_width = int(min(width, height * (3 / 4)))
        target_height = int(min(height, width * (4 / 3)))        
        
        left = (width - target_width) / 2
        right = (width + target_width) / 2
        # for Landmark, model sizes are 594x879, so we need to reduce the height. In some case the garment on the model is
        # also getting removed when reducing size from bottom. So we will only reduce height from top for now
        top = (height - target_height) #top = (height - target_height) / 2        
        bottom = height #bottom = (height + target_height) / 2
        cropped_img = human_img_orig.crop((left, top, right, bottom))
        
        crop_size = cropped_img.size
        human_img = cropped_img.resize((WIDTH, HEIGHT))
    else:
        human_img = human_img_orig.resize((WIDTH, HEIGHT))

    # Commenting out naize harmonization for now. We will have to integrate with Deep Learning based Harmonization methods
    # Do color transfer from background image for better image harmonization
    #if background_img:
    #    human_img = BackgroundProcessor.intensity_transfer(human_img, background_img)


    if is_checked:
        # internally openpose_model is resizing human_img to resolution 384 if not passed as input
        keypoints = openpose_model(human_img.resize((POSE_WIDTH, POSE_HEIGHT)))
        model_parse, _ = parsing_model(human_img.resize((POSE_WIDTH, POSE_HEIGHT)))
        # internally get mask location function is resizing model_parse to 384x512 if width & height not passed
        mask, mask_gray = get_mask_location(ARM_WIDTH, CATEGORY, model_parse, keypoints)
        mask = mask.resize((WIDTH, HEIGHT))
        logging.info("Mask location on model identified")
    else:
        mask = pil_to_binary_mask(human_img_dict['layers'][0].convert("RGB").resize((WIDTH, HEIGHT)))
        # mask = transforms.ToTensor()(mask)
        # mask = mask.unsqueeze(0)
    mask_gray = (1-transforms.ToTensor()(mask)) * tensor_transfrom(human_img)
    mask_gray = to_pil_image((mask_gray+1.0)/2.0)


    human_img_arg = _apply_exif_orientation(human_img.resize((POSE_WIDTH,POSE_HEIGHT)))
    human_img_arg = convert_PIL_to_numpy(human_img_arg, format="BGR")
     
    

    args = apply_net.create_argument_parser().parse_args(('show', './configs/densepose_rcnn_R_50_FPN_s1x.yaml', './ckpt/densepose/model_final_162be9.pkl', 'dp_segm', '-v', '--opts', 'MODEL.DEVICE', device))
    # verbosity = getattr(args, "verbosity", None)
    pose_img = args.func(args,human_img_arg)    
    pose_img = pose_img[:,:,::-1]    
    pose_img = Image.fromarray(pose_img).resize((WIDTH,HEIGHT))
    
    with torch.no_grad():
        # Extract the images
        with torch.cuda.amp.autocast():
            with torch.no_grad():
                prompt = "model is wearing " + garment_des
                negative_prompt = "monochrome, lowres, bad anatomy, worst quality, low quality"
                with torch.inference_mode():
                    (
                        prompt_embeds,
                        negative_prompt_embeds,
                        pooled_prompt_embeds,
                        negative_pooled_prompt_embeds,
                    ) = pipe.encode_prompt(
                        prompt,
                        num_images_per_prompt=1,
                        do_classifier_free_guidance=True,
                        negative_prompt=negative_prompt,
                    )
                                    
                    prompt = "a photo of " + garment_des
                    negative_prompt = "monochrome, lowres, bad anatomy, worst quality, low quality"
                    if not isinstance(prompt, List):
                        prompt = [prompt] * 1
                    if not isinstance(negative_prompt, List):
                        negative_prompt = [negative_prompt] * 1
                    with torch.inference_mode():
                        (
                            prompt_embeds_c,
                            _,
                            _,
                            _,
                        ) = pipe.encode_prompt(
                            prompt,
                            num_images_per_prompt=1,
                            do_classifier_free_guidance=False,
                            negative_prompt=negative_prompt,
                        )



                    pose_img =  tensor_transfrom(pose_img).unsqueeze(0).to(device,torch.float16)
                    garm_tensor =  tensor_transfrom(garm_img).unsqueeze(0).to(device,torch.float16)
                    generator = torch.Generator(device).manual_seed(seed) if seed is not None else None
                    images = pipe(
                        prompt_embeds=prompt_embeds.to(device,torch.float16),
                        negative_prompt_embeds=negative_prompt_embeds.to(device,torch.float16),
                        pooled_prompt_embeds=pooled_prompt_embeds.to(device,torch.float16),
                        negative_pooled_prompt_embeds=negative_pooled_prompt_embeds.to(device,torch.float16),
                        num_inference_steps=denoise_steps,
                        generator=generator,
                        strength = 1.0,
                        pose_img = pose_img.to(device,torch.float16),
                        text_embeds_cloth=prompt_embeds_c.to(device,torch.float16),
                        cloth = garm_tensor.to(device,torch.float16),
                        mask_image=mask,
                        image=human_img, 
                        height=HEIGHT,
                        width=WIDTH,
                        ip_adapter_image = garm_img.resize((WIDTH,HEIGHT)),
                        guidance_scale=2.0,
                    )[0]

    if is_checked_crop:
        out_img = images[0].resize(crop_size)        
        human_img_orig.paste(out_img, (int(left), int(top)))    
        final_image = human_img_orig
        # return human_img_orig, mask_gray
    else:
        final_image = images[0]
        # return images[0], mask_gray
    
    # apply background to final image
    if background_img:
       logging.info("Adding background")
       final_image = BackgroundProcessor.replace_background_with_removebg(final_image, background_img) 
    return final_image, mask_gray
    # return images[0], mask_gray

garm_list = os.listdir(os.path.join(example_path,"cloth"))
garm_list_path = [os.path.join(example_path,"cloth",garm) for garm in garm_list]

human_list = os.listdir(os.path.join(example_path,"human"))
human_list_path = [os.path.join(example_path,"human",human) for human in human_list]

human_ex_list = []
human_ex_list = human_list_path # Image
""" if using ImageEditor instead of Image while taking input, use this - ImageEditor
for ex_human in human_list_path:
    ex_dict= {}
    ex_dict['background'] = ex_human
    ex_dict['layers'] = None
    ex_dict['composite'] = None
    human_ex_list.append(ex_dict)
"""
##default human


# api_open=True will allow this API to be hit using curl
image_blocks = gr.Blocks().queue(api_open=True)
with image_blocks as demo:
    gr.Markdown("## Virtual Try-On 👕👔👚")
    gr.Markdown("Upload an image of a person and an image of a garment ✨.")
    with gr.Row():
        with gr.Column():
            # changing from ImageEditor to Image to allow easy passing of data through API
            # instead of passing {"dictionary": <>} ( which is failing ), we can directly pass the image
            #imgs = gr.ImageEditor(sources='upload', type="pil", label='Human. Mask with pen or use auto-masking', interactive=True)
            imgs = gr.Image(sources='upload', type='pil',label='Human. Mask with pen or use auto-masking')
            with gr.Row():
                is_checked = gr.Checkbox(label="Yes", info="Use auto-generated mask (Takes 5 seconds)",value=True)
            with gr.Row():
                is_checked_crop = gr.Checkbox(label="Yes", info="Use auto-crop & resizing",value=False)

            example = gr.Examples(
                inputs=imgs,
                examples_per_page=10,
                examples=human_ex_list
            )

        with gr.Column():
            garm_img = gr.Image(label="Garment", sources='upload', type="pil")
            with gr.Row(elem_id="prompt-container"):
                with gr.Row():
                    prompt = gr.Textbox(placeholder="Description of garment ex) Short Sleeve Round Neck T-shirts", show_label=False, elem_id="prompt")
            example = gr.Examples(
                inputs=garm_img,
                examples_per_page=8,
                examples=garm_list_path)
            
        with gr.Column():
           background_img = gr.Image(label="Background", sources='upload', type="pil")

        with gr.Column():
            with gr.Row():
               image_out = gr.Image(label="Output", elem_id="output-img", show_share_button=False) 
            with gr.Row():
               masked_img = gr.Image(label="Masked image output", elem_id="masked-img", show_share_button=False) 
        """
        with gr.Column():
            # image_out = gr.Image(label="Output", elem_id="output-img", height=400)
            masked_img = gr.Image(label="Masked image output", elem_id="masked-img", show_share_button=False)
        with gr.Column():            
            # image_out = gr.Image(label="Output", elem_id="output-img", height=400)
            image_out = gr.Image(label="Output", elem_id="output-img", show_share_button=False)
        """



    with gr.Column():
        try_button = gr.Button(value="Try-on")
        with gr.Accordion(label="Advanced Settings", open=False):
            with gr.Row():
                denoise_steps = gr.Number(label="Denoising Steps", minimum=20, maximum=40, value=30, step=1)
                seed = gr.Number(label="Seed", minimum=-1, maximum=2147483647, step=1, value=42)



    try_button.click(fn=start_tryon, inputs=[imgs, garm_img, prompt, background_img, is_checked,is_checked_crop, denoise_steps, seed], outputs=[image_out,masked_img], api_name='tryon')

            
image_blocks.launch()