app_d.py

import gradio as gr
import os
import cv2
import numpy as np
from moviepy.editor import *
#from share_btn import community_icon_html, loading_icon_html, share_js


os.system("python -m pip install git+https://github.com/MaureenZOU/detectron2-xyz.git")


import torch
import argparse

from xdecoder.BaseModel import BaseModel
from xdecoder import build_model
from utils.distributed import init_distributed
from utils.arguments import load_opt_from_config_files

from tasks import *

def parse_option():
    parser = argparse.ArgumentParser('X-Decoder All-in-One Demo', add_help=False)
    parser.add_argument('--conf_files', default="configs/xdecoder/svlp_focalt_lang.yaml", metavar="FILE", help='path to config file', )
    args = parser.parse_args()

    return args

'''
build args
'''
args = parse_option()
opt = load_opt_from_config_files(args.conf_files)
opt = init_distributed(opt)

# META DATA
pretrained_pth_last = os.path.join("xdecoder_focalt_last.pt")
pretrained_pth_novg = os.path.join("xdecoder_focalt_last_novg.pt")

if not os.path.exists(pretrained_pth_last):
    os.system("wget {}".format("https://projects4jw.blob.core.windows.net/x-decoder/release/xdecoder_focalt_last.pt"))

if not os.path.exists(pretrained_pth_novg):
    os.system("wget {}".format("https://projects4jw.blob.core.windows.net/x-decoder/release/xdecoder_focalt_last_novg.pt"))


'''
build model
'''
model_last = BaseModel(opt, build_model(opt)).from_pretrained(pretrained_pth_last).eval().cuda()

with torch.no_grad():
    model_last.model.sem_seg_head.predictor.lang_encoder.get_text_embeddings(["background", "background"], is_eval=True)

'''
inference model
'''

@torch.no_grad()
def xdecoder(image, instruction, *args, **kwargs):
    image = image.convert("RGB")
    with torch.autocast(device_type='cuda', dtype=torch.float16):
        return referring_inpainting_gpt3(model_last, image, instruction, *args, **kwargs)


#xdecoder = gr.Interface.load(name="spaces/xdecoder/Instruct-X-Decoder")

def get_frames(video_in):
    frames = []
    #resize the video
    clip = VideoFileClip(video_in)
    
    #check fps
    if clip.fps > 30:
        print("vide rate is over 30, resetting to 30")
        clip_resized = clip.resize(height=512)
        clip_resized.write_videofile("video_resized.mp4", fps=30)
    else:
        print("video rate is OK")
        clip_resized = clip.resize(height=512)
        clip_resized.write_videofile("video_resized.mp4", fps=clip.fps)
    
    print("video resized to 512 height")
    
    # Opens the Video file with CV2
    cap= cv2.VideoCapture("video_resized.mp4")
    
    fps = cap.get(cv2.CAP_PROP_FPS)
    print("video fps: " + str(fps))
    i=0
    while(cap.isOpened()):
        ret, frame = cap.read()
        if ret == False:
            break
        cv2.imwrite('kang'+str(i)+'.jpg',frame)
        frames.append('kang'+str(i)+'.jpg')
        i+=1
    
    cap.release()
    cv2.destroyAllWindows()
    print("broke the video into frames")
    
    return frames, fps


def create_video(frames, fps):
    print("building video result")
    clip = ImageSequenceClip(frames, fps=fps)
    clip.write_videofile("movie.mp4", fps=fps)
    
    return 'movie.mp4'


def infer(prompt,video_in, trim_value):
    print(prompt)
    break_vid = get_frames(video_in)
    
    frames_list= break_vid[0]
    fps = break_vid[1]
    n_frame = int(trim_value*fps)
    
    if n_frame >= len(frames_list):
        print("video is shorter than the cut value")
        n_frame = len(frames_list)
    
    result_frames = []
    print("set stop frames to: " + str(n_frame))
    
    for i in frames_list[0:int(n_frame)]:
        #xdecoder_img = xdecoder(i, prompt, fn_index=0)
        xdecoder_img = xdecoder(i, prompt)
        #res_image = xdecoder_img[0]
        #rgb_im = images[0].convert("RGB")
  
        # exporting the image
        #res_image.save(f"result_img-{i}.jpg")
        result_frames.append(xdecoder_img)
        print("frame " + i + "/" + str(n_frame) + ": done;")

    print(result_frames)
    final_vid = create_video(result_frames, fps)
    print("finished !")
    
    #return final_vid, gr.Group.update(visible=True)
    return final_vid

title = """
    <div style="text-align: center; max-width: 700px; margin: 0 auto;">
        <div
        style="
            display: inline-flex;
            align-items: center;
            gap: 0.8rem;
            font-size: 1.75rem;
        "
        >
        <h1 style="font-weight: 900; margin-bottom: 7px; margin-top: 5px;">
            Instruct X-Decoder Video
        </h1>
        </div>
        <p style="margin-bottom: 10px; font-size: 94%">
        Apply Instruct X-Decoder Diffusion to a video 
        </p>
    </div>
"""

article = """
    
    <div class="footer">
        <p>
        Examples by <a href="https://twitter.com/CitizenPlain" target="_blank">Nathan Shipley</a> •&nbsp;
        Follow <a href="https://twitter.com/fffiloni" target="_blank">Sylvain Filoni</a> for future updates 🤗
        </p>
    </div>
    <div id="may-like-container" style="display: flex;justify-content: center;flex-direction: column;align-items: center;margin-bottom: 30px;">
        <p>You may also like: </p>
        <div id="may-like-content" style="display:flex;flex-wrap: wrap;align-items:center;height:20px;">
            
            <svg height="20" width="162" style="margin-left:4px;margin-bottom: 6px;">       
                 <a href="https://huggingface.co./spaces/timbrooks/instruct-pix2pix" target="_blank">
                    <image href="https://img.shields.io/badge/🤗 Spaces-Instruct_Pix2Pix-blue" src="https://img.shields.io/badge/🤗 Spaces-Instruct_Pix2Pix-blue.png" height="20"/>
                 </a>
            </svg>
            
        </div>
    
    </div>
    
"""

with gr.Blocks(css='style.css') as demo:
    with gr.Column(elem_id="col-container"):
        gr.HTML(title)
        with gr.Row():
            with gr.Column():
                video_inp = gr.Video(label="Video source", source="upload", type="filepath", elem_id="input-vid")
                prompt = gr.Textbox(label="Prompt", placeholder="enter prompt", show_label=False, elem_id="prompt-in")
                with gr.Row():
                    trim_in = gr.Slider(label="Cut video at (s)", minimun=1, maximum=3, step=1, value=1)
            with gr.Column():
                video_out = gr.Video(label="Pix2pix video result", elem_id="video-output")
                gr.HTML("""
                <a style="display:inline-block" href="https://huggingface.co./spaces/fffiloni/Pix2Pix-Video?duplicate=true"><img src="https://img.shields.io/badge/-Duplicate%20Space-blue?labelColor=white&style=flat&logo=data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/9hAAAAAXNSR0IArs4c6QAAAP5JREFUOE+lk7FqAkEURY+ltunEgFXS2sZGIbXfEPdLlnxJyDdYB62sbbUKpLbVNhyYFzbrrA74YJlh9r079973psed0cvUD4A+4HoCjsA85X0Dfn/RBLBgBDxnQPfAEJgBY+A9gALA4tcbamSzS4xq4FOQAJgCDwV2CPKV8tZAJcAjMMkUe1vX+U+SMhfAJEHasQIWmXNN3abzDwHUrgcRGmYcgKe0bxrblHEB4E/pndMazNpSZGcsZdBlYJcEL9Afo75molJyM2FxmPgmgPqlWNLGfwZGG6UiyEvLzHYDmoPkDDiNm9JR9uboiONcBXrpY1qmgs21x1QwyZcpvxt9NS09PlsPAAAAAElFTkSuQmCC&logoWidth=14" alt="Duplicate Space"></a> 
                work with longer videos / skip the queue: 
                """, elem_id="duplicate-container")
                submit_btn = gr.Button("Generate X-Decoder video")

                #with gr.Group(elem_id="share-btn-container", visible=False) as share_group:
                #    community_icon = gr.HTML(community_icon_html)
                #    loading_icon = gr.HTML(loading_icon_html)
                #    share_button = gr.Button("Share to community", elem_id="share-btn")
        
        inputs = [prompt, video_inp, trim_in]
        #outputs = [video_out, share_group]
        outputs = [video_out]
        
        gr.HTML(article)
    
    submit_btn.click(infer, inputs, outputs)
    #share_button.click(None, [], [], _js=share_js)

    
    
demo.launch().queue(max_size=12)