gradio_app.py

import logging
import os
import tempfile
import time

import gradio as gr
import numpy as np
import rembg
import torch
from PIL import Image
from functools import partial

from tsr.system import TSR
from tsr.utils import remove_background, resize_foreground, to_gradio_3d_orientation

from diffusers import ControlNetModel, StableDiffusionControlNetPipeline, DiffusionPipeline
from diffusion_webui.diffusion_models.base_controlnet_pipeline import (
    ControlnetPipeline,
)
from diffusion_webui.utils.model_list import (
    controlnet_model_list,
    stable_model_list,
)
from diffusion_webui.utils.preprocces_utils import PREPROCCES_DICT
from diffusion_webui.utils.scheduler_list import (
    SCHEDULER_MAPPING,
    get_scheduler,
)
if torch.cuda.is_available():
    device = "cuda:0"
else:
    device = "cpu"



import argparse


if torch.cuda.is_available():
    device = "cuda:0"
else:
    device = "cpu"

model = TSR.from_pretrained(
    "stabilityai/TripoSR",
    config_name="config.yaml",
    weight_name="model.ckpt",
)

# adjust the chunk size to balance between speed and memory usage
model.renderer.set_chunk_size(8192)
model.to(device)

rembg_session = rembg.new_session()


def check_input_image(input_image):
    if input_image is None:
        raise gr.Error("No image uploaded!")


def preprocess(input_image, do_remove_background, foreground_ratio):
    def fill_background(image):
        image = np.array(image).astype(np.float32) / 255.0
        image = image[:, :, :3] * image[:, :, 3:4] + (1 - image[:, :, 3:4]) * 0.5
        image = Image.fromarray((image * 255.0).astype(np.uint8))
        return image

    if do_remove_background:
        image = input_image.convert("RGB")
        image = remove_background(image, rembg_session)
        image = resize_foreground(image, foreground_ratio)
        image = fill_background(image)
    else:
        image = input_image
        if image.mode == "RGBA":
            image = fill_background(image)
    return image


def generate(image, mc_resolution, formats=["obj", "glb"]):
    scene_codes = model(image, device=device)
    mesh = model.extract_mesh(scene_codes, resolution=mc_resolution)[0]
    mesh = to_gradio_3d_orientation(mesh)
    rv = []
    for format in formats:
        mesh_path = tempfile.NamedTemporaryFile(suffix=f".{format}", delete=False)
        mesh.export(mesh_path.name)
        rv.append(mesh_path.name)
    return rv[0]


class StableDiffusionControlNetGenerator(ControlnetPipeline):
    def __init__(self):
        self.pipe = None

    def load_model(self, stable_model_path, controlnet_model_path, scheduler):
        if self.pipe is None:
            controlnet = ControlNetModel.from_pretrained(
                controlnet_model_path, torch_dtype=torch.float16, cache_dir="./model/"
            )
            self.pipe = StableDiffusionControlNetPipeline.from_pretrained(
                pretrained_model_name_or_path=stable_model_path,
                controlnet=controlnet,
                safety_checker=None,
                torch_dtype=torch.float16,
                cache_dir="./model/"
            )

        self.pipe = get_scheduler(pipe=self.pipe, scheduler=scheduler)
        self.pipe.to("cuda")
        self.pipe.enable_xformers_memory_efficient_attention()

        return self.pipe

    def controlnet_preprocces(
            self,
            read_image: str,
            preprocces_type: str,
    ):
        processed_image = PREPROCCES_DICT[preprocces_type](read_image)
        return processed_image

    def generate_image(
            self,
            image_path: str,
            stable_model_path: str,
            controlnet_model_path: str,
            height: int,
            width: int,
            guess_mode: bool,
            controlnet_conditioning_scale: int,
            prompt: str,
            negative_prompt: str,
            num_images_per_prompt: int,
            guidance_scale: int,
            num_inference_step: int,
            scheduler: str,
            seed_generator: int,
            preprocces_type: str = "Lineart",
    ):
        print("stable model: ", stable_model_path)
        print("controlnet model: ", controlnet_model_path)
        pipe = self.load_model(
            stable_model_path=stable_model_path,
            controlnet_model_path=controlnet_model_path,
            scheduler=scheduler,
        )

        read_image = image_path
        controlnet_image = self.controlnet_preprocces(
            read_image=read_image, preprocces_type=preprocces_type
        )

        if seed_generator == 0:
            random_seed = torch.randint(0, 1000000, (1,))
            generator = torch.manual_seed(random_seed)
        else:
            generator = torch.manual_seed(seed_generator)

        output = pipe(
            prompt=prompt,
            height=height,
            width=width,
            controlnet_conditioning_scale=float(controlnet_conditioning_scale),
            guess_mode=guess_mode,
            image=controlnet_image,
            negative_prompt=negative_prompt,
            num_images_per_prompt=num_images_per_prompt,
            num_inference_steps=num_inference_step,
            guidance_scale=guidance_scale,
            generator=generator,
        ).images
        return output[0]

def get_sketch(im):
    return im["composite"]

with gr.Blocks(title="scratch to 3D") as interface:
    # gr.Markdown(
    #     """
    # # TripoSR Demo
    # [TripoSR](https://github.com/VAST-AI-Research/TripoSR) is a state-of-the-art open-source model for **fast** feedforward 3D reconstruction from a single image, collaboratively developed by [Tripo AI](https://www.tripo3d.ai/) and [Stability AI](https://stability.ai/).
    #
    # **Tips:**
    # 1. If you find the result is unsatisfied, please try to change the foreground ratio. It might improve the results.
    # 2. It's better to disable "Remove Background" for the provided examples (except fot the last one) since they have been already preprocessed.
    # 3. Otherwise, please disable "Remove Background" option only if your input image is RGBA with transparent background, image contents are centered and occupy more than 70% of image width or height.
    # """
    # )
    with gr.Row(variant="panel"):
        with gr.Column():
            controlnet_image_path = gr.ImageEditor(
                sources="upload",
                label="Input Image",
                type="pil",
                # height=700,
            )

        with gr.Column():
            controlnet_prompt = gr.Textbox(
                lines=1, placeholder="Prompt", show_label=False
            )
            controlnet_negative_prompt = gr.Textbox(
                lines=1, placeholder="Negative Prompt", show_label=False
            )

            with gr.Row(visible=False):
                with gr.Column():
                    controlnet_stable_model_path = gr.Dropdown(
                        choices=stable_model_list,
                        value=stable_model_list[0],
                        label="Stable Model Path",
                        visible=True,
                        interactive=True
                    )
                    controlnet_preprocces_type = gr.Dropdown(
                        choices=list(PREPROCCES_DICT.keys()),
                        value="Lineart",
                        label="Preprocess Type",
                        visible=False,
                        interactive=True
                    )
                    controlnet_conditioning_scale = gr.Slider(
                        minimum=0.0,
                        maximum=1.0,
                        step=0.1,
                        value=1.0,
                        visible=True,
                        label="ControlNet Conditioning Scale",
                        interactive=True
                    )
                    controlnet_guidance_scale = gr.Slider(
                        minimum=0.1,
                        maximum=15,
                        step=0.1,
                        value=7.5,
                        visible=True,
                        label="Guidance Scale",
                        interactive=True
                    )

                    controlnet_width = gr.Slider(
                        minimum=128,
                        maximum=1280,
                        step=32,
                        value=512,
                        visible=False,
                        label="Width",
                        interactive=True
                    )

                with gr.Column():
                    controlnet_model_path = gr.Dropdown(
                        choices=controlnet_model_list,
                        value="lllyasviel/control_v11p_sd15_lineart",
                        label="ControlNet Model Path",
                        visible=True,
                        interactive=True
                    )
                    controlnet_scheduler = gr.Dropdown(
                        choices=list(SCHEDULER_MAPPING.keys()),
                        value=list(SCHEDULER_MAPPING.keys())[0],
                        label="Scheduler",
                        visible=True,
                        interactive=True
                    )
                    controlnet_num_inference_step = gr.Slider(
                        minimum=1,
                        maximum=150,
                        step=1,
                        value=30,
                        label="Num Inference Step",
                        visible=True,
                        interactive=True
                    )
                    controlnet_num_images_per_prompt = gr.Slider(
                        minimum=1,
                        maximum=4,
                        step=1,
                        value=1,
                        label="Number Of Images",
                        visible=False,
                        interactive=True
                    )
                    controlnet_seed_generator = gr.Slider(
                        minimum=0,
                        maximum=1000000,
                        step=1,
                        value=0,
                        label="Seed(0 for random)",
                        visible=False,
                        interactive=True
                    )
                    controlnet_guess_mode = gr.Checkbox(
                        label="Guess Mode",
                        value=True,
                        visible=False,
                        interactive=True
                    )
                    controlnet_height = gr.Slider(
                        minimum=128,
                        maximum=1280,
                        step=32,
                        value=512,
                        label="Height",
                        visible=False,
                        interactive=True
                    )

            with gr.Row():
                with gr.Column():
                    with gr.Row():
                        input_image = gr.Image(
                            label="2D Image",
                            image_mode="RGBA",
                            type="pil",
                            elem_id="content_image",
                            height=400,
                            interactive=False,

                        )
                        processed_image = gr.Image(
                            label="Processed Image",
                            interactive=False,
                            height=400,
                            visible=False
                        )
                    with gr.Row():
                        generate_2d = gr.Button("Generate 2D", elem_id="generate", variant="primary")

                with gr.Column():
                    with gr.Row():
                        output_model_obj = gr.Model3D(
                            label="Output Model (OBJ Format)",
                            interactive=False,
                            height=400,
                        )
                    with gr.Row():
                        generate_3d = gr.Button("Generate 3D", elem_id="generate", variant="primary")

    with gr.Row(visible=False):
        with gr.Group():
            do_remove_background = gr.Checkbox(
                label="Remove Background", value=True
            )
            foreground_ratio = gr.Slider(
                label="Foreground Ratio",
                minimum=0.5,
                maximum=1.0,
                value=0.85,
                step=0.05,
            )
            mc_resolution = gr.Slider(
                label="Marching Cubes Resolution",
                minimum=32,
                maximum=320,
                value=256,
                step=32
            )


    generate_2d.click(
        fn=get_sketch,
        inputs=[controlnet_image_path],
        outputs=[input_image]
    ).success(
        fn=StableDiffusionControlNetGenerator().generate_image,
        inputs=[
            input_image,
            controlnet_stable_model_path,
            controlnet_model_path,
            controlnet_height,
            controlnet_width,
            controlnet_guess_mode,
            controlnet_conditioning_scale,
            controlnet_prompt,
            controlnet_negative_prompt,
            controlnet_num_images_per_prompt,
            controlnet_guidance_scale,
            controlnet_num_inference_step,
            controlnet_scheduler,
            controlnet_seed_generator,
            controlnet_preprocces_type,
        ],
        outputs=[input_image],
    ).success(
        fn=preprocess,
        inputs=[input_image, do_remove_background, foreground_ratio],
        outputs=[processed_image],
    )

    generate_3d.click(fn=check_input_image, inputs=[processed_image]).success(
        fn=generate,
        inputs=[processed_image, mc_resolution],
        outputs=[output_model_obj],
    )

if __name__ == '__main__':

    interface.launch(server_name='0.0.0.0', share=False)