Upload 425 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+temp/temp/temp.wav filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -0,0 +1,28 @@
+# Use Python 3.9 base image
+FROM python:3.9
+
+# Set the working directory
+WORKDIR /code
+
+# Copy requirements file and install dependencies
+COPY ./requirements.txt /code/requirements.txt
+RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
+
+# Copy the rest of the project files
+COPY ./models /code/models
+COPY ./checkpoints /code/checkpoints
+COPY ./results /code/results
+COPY ./temp /code/temp
+COPY ./third_part /code/third_part
+COPY ./utils /code/utils
+
+# Copy additional Python files
+COPY ./inference.py /code/inference.py
+COPY ./predict.py /code/predict.py
+COPY ./app.py /code/app.py
+
+# Expose port for FastAPI server
+EXPOSE 7860
+
+# Define the command to run the FastAPI application with uvicorn
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -0,0 +1,58 @@
+import os
+import shutil
+import subprocess
+import tempfile
+import urllib.request
+from typing import Tuple, Optional
+
+import asyncio
+
+from fastapi import FastAPI, File, UploadFile, HTTPException
+from fastapi.responses import FileResponse
+from pydantic import BaseModel, HttpUrl
+
+app = FastAPI()
+
+class VideoAudioRequest(BaseModel):
+    face_url: HttpUrl
+    audio_url: HttpUrl
+
+def download_file(url: str, destination: str) -> None:
+    try:
+        with urllib.request.urlopen(str(url)) as response, open(destination, 'wb') as out_file:
+            shutil.copyfileobj(response, out_file)
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to download file from {url}: {e}")
+
+async def process_video_and_audio(face_url: str, audio_url: str) -> Tuple[str, Optional[str]]:
+    print("process started")
+    temp_dir = tempfile.mkdtemp(prefix="fastapi_processing_")
+    face_path = os.path.join(temp_dir, "face.mp4")
+    audio_path = os.path.join(temp_dir, "audio.mp3")
+
+    try:
+        download_file(face_url, face_path)
+        download_file(audio_url, audio_path)
+        print(face_path,audio_path)
+    except HTTPException as e:
+        return "", str(e)
+
+    outfile_path = os.path.join(temp_dir, "result.mp4")
+    command = f"python3 inference.py --face {face_path} --audio {audio_path} --outfile {outfile_path}"
+    try:
+        process = await asyncio.create_subprocess_shell(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        _, error = await process.communicate()
+        if process.returncode == 0:
+            return outfile_path, None
+        else:
+            return "", f"Error occurred during processing: {error.decode()}"
+    except Exception as e:
+        return "", f"Error occurred during processing: {e}"
+
+@app.post("/process_video_audio")
+async def process_video_audio(request_data: VideoAudioRequest):
+    result, error = await process_video_and_audio(request_data.face_url, request_data.audio_url)
+    if result:
+        return FileResponse(result, media_type="video/mp4")
+    else:
+        raise HTTPException(status_code=500, detail=error)

inference.py

@@ -0,0 +1,345 @@
+import numpy as np
+import cv2, os, sys, subprocess, platform, torch
+from tqdm import tqdm
+from PIL import Image
+from scipy.io import loadmat
+
+sys.path.insert(0, 'third_part')
+sys.path.insert(0, 'third_part/GPEN')
+sys.path.insert(0, 'third_part/GFPGAN')
+
+# 3dmm extraction
+from third_part.face3d.util.preprocess import align_img
+from third_part.face3d.util.load_mats import load_lm3d
+from third_part.face3d.extract_kp_videos import KeypointExtractor
+# face enhancement
+from third_part.GPEN.gpen_face_enhancer import FaceEnhancement
+from third_part.GFPGAN.gfpgan import GFPGANer
+# expression control
+from third_part.ganimation_replicate.model.ganimation import GANimationModel
+
+from utils import audio
+from utils.ffhq_preprocess import Croper
+from utils.alignment_stit import crop_faces, calc_alignment_coefficients, paste_image
+from utils.inference_utils import Laplacian_Pyramid_Blending_with_mask, face_detect, load_model, options, split_coeff, \
+                                  trans_image, transform_semantic, find_crop_norm_ratio, load_face3d_net, exp_aus_dict
+import warnings
+warnings.filterwarnings("ignore")
+
+args = options()
+
+def main():    
+    device = 'cuda' if torch.cuda.is_available() else 'cpu'
+    print('[Info] Using {} for inference.'.format(device))
+    os.makedirs(os.path.join('temp', args.tmp_dir), exist_ok=True)
+
+    enhancer = FaceEnhancement(base_dir='checkpoints', size=512, model='GPEN-BFR-512', use_sr=False, \
+                               sr_model='rrdb_realesrnet_psnr', channel_multiplier=2, narrow=1, device=device)
+    restorer = GFPGANer(model_path='checkpoints/GFPGANv1.3.pth', upscale=1, arch='clean', \
+                        channel_multiplier=2, bg_upsampler=None)
+
+    base_name = args.face.split('/')[-1]
+    if os.path.isfile(args.face) and args.face.split('.')[1] in ['jpg', 'png', 'jpeg']:
+        args.static = True
+    if not os.path.isfile(args.face):
+        raise ValueError('--face argument must be a valid path to video/image file')
+    elif args.face.split('.')[1] in ['jpg', 'png', 'jpeg']:
+        full_frames = [cv2.imread(args.face)]
+        fps = args.fps
+    else:
+        video_stream = cv2.VideoCapture(args.face)
+        fps = video_stream.get(cv2.CAP_PROP_FPS)
+
+        full_frames = []
+        while True:
+            still_reading, frame = video_stream.read()
+            if not still_reading:
+                video_stream.release()
+                break
+            y1, y2, x1, x2 = args.crop
+            if x2 == -1: x2 = frame.shape[1]
+            if y2 == -1: y2 = frame.shape[0]
+            frame = frame[y1:y2, x1:x2]
+            full_frames.append(frame)
+
+    print ("[Step 0] Number of frames available for inference: "+str(len(full_frames)))
+    # face detection & cropping, cropping the first frame as the style of FFHQ
+    croper = Croper('checkpoints/shape_predictor_68_face_landmarks.dat')
+    full_frames_RGB = [cv2.cvtColor(frame, cv2.COLOR_BGR2RGB) for frame in full_frames]
+    full_frames_RGB, crop, quad = croper.crop(full_frames_RGB, xsize=512)
+
+    clx, cly, crx, cry = crop
+    lx, ly, rx, ry = quad
+    lx, ly, rx, ry = int(lx), int(ly), int(rx), int(ry)
+    oy1, oy2, ox1, ox2 = cly+ly, min(cly+ry, full_frames[0].shape[0]), clx+lx, min(clx+rx, full_frames[0].shape[1])
+    # original_size = (ox2 - ox1, oy2 - oy1)
+    frames_pil = [Image.fromarray(cv2.resize(frame,(256,256))) for frame in full_frames_RGB]
+
+    # get the landmark according to the detected face.
+    if not os.path.isfile('temp/'+base_name+'_landmarks.txt') or args.re_preprocess:
+        print('[Step 1] Landmarks Extraction in Video.')
+        kp_extractor = KeypointExtractor()
+        lm = kp_extractor.extract_keypoint(frames_pil, './temp/'+base_name+'_landmarks.txt')
+    else:
+        print('[Step 1] Using saved landmarks.')
+        lm = np.loadtxt('temp/'+base_name+'_landmarks.txt').astype(np.float32)
+        lm = lm.reshape([len(full_frames), -1, 2])
+       
+    if not os.path.isfile('temp/'+base_name+'_coeffs.npy') or args.exp_img is not None or args.re_preprocess:
+        net_recon = load_face3d_net(args.face3d_net_path, device)
+        lm3d_std = load_lm3d('checkpoints/BFM')
+
+        video_coeffs = []
+        for idx in tqdm(range(len(frames_pil)), desc="[Step 2] 3DMM Extraction In Video:"):
+            frame = frames_pil[idx]
+            W, H = frame.size
+            lm_idx = lm[idx].reshape([-1, 2])
+            if np.mean(lm_idx) == -1:
+                lm_idx = (lm3d_std[:, :2]+1) / 2.
+                lm_idx = np.concatenate([lm_idx[:, :1] * W, lm_idx[:, 1:2] * H], 1)
+            else:
+                lm_idx[:, -1] = H - 1 - lm_idx[:, -1]
+
+            trans_params, im_idx, lm_idx, _ = align_img(frame, lm_idx, lm3d_std)
+            trans_params = np.array([float(item) for item in np.hsplit(trans_params, 5)]).astype(np.float32)
+            im_idx_tensor = torch.tensor(np.array(im_idx)/255., dtype=torch.float32).permute(2, 0, 1).to(device).unsqueeze(0) 
+            with torch.no_grad():
+                coeffs = split_coeff(net_recon(im_idx_tensor))
+
+            pred_coeff = {key:coeffs[key].cpu().numpy() for key in coeffs}
+            pred_coeff = np.concatenate([pred_coeff['id'], pred_coeff['exp'], pred_coeff['tex'], pred_coeff['angle'],\
+                                         pred_coeff['gamma'], pred_coeff['trans'], trans_params[None]], 1)
+            video_coeffs.append(pred_coeff)
+        semantic_npy = np.array(video_coeffs)[:,0]
+        np.save('temp/'+base_name+'_coeffs.npy', semantic_npy)
+    else:
+        print('[Step 2] Using saved coeffs.')
+        semantic_npy = np.load('temp/'+base_name+'_coeffs.npy').astype(np.float32)
+
+    # generate the 3dmm coeff from a single image
+    if args.exp_img is not None and ('.png' in args.exp_img or '.jpg' in args.exp_img):
+        print('extract the exp from',args.exp_img)
+        exp_pil = Image.open(args.exp_img).convert('RGB')
+        lm3d_std = load_lm3d('third_part/face3d/BFM')
+        
+        W, H = exp_pil.size
+        kp_extractor = KeypointExtractor()
+        lm_exp = kp_extractor.extract_keypoint([exp_pil], 'temp/'+base_name+'_temp.txt')[0]
+        if np.mean(lm_exp) == -1:
+            lm_exp = (lm3d_std[:, :2] + 1) / 2.
+            lm_exp = np.concatenate(
+                [lm_exp[:, :1] * W, lm_exp[:, 1:2] * H], 1)
+        else:
+            lm_exp[:, -1] = H - 1 - lm_exp[:, -1]
+
+        trans_params, im_exp, lm_exp, _ = align_img(exp_pil, lm_exp, lm3d_std)
+        trans_params = np.array([float(item) for item in np.hsplit(trans_params, 5)]).astype(np.float32)
+        im_exp_tensor = torch.tensor(np.array(im_exp)/255., dtype=torch.float32).permute(2, 0, 1).to(device).unsqueeze(0)
+        with torch.no_grad():
+            expression = split_coeff(net_recon(im_exp_tensor))['exp'][0]
+        del net_recon
+    elif args.exp_img == 'smile':
+        expression = torch.tensor(loadmat('checkpoints/expression.mat')['expression_mouth'])[0]
+    else:
+        print('using expression center')
+        expression = torch.tensor(loadmat('checkpoints/expression.mat')['expression_center'])[0]
+
+    # load DNet, model(LNet and ENet)
+    D_Net, model = load_model(args, device)
+
+    if not os.path.isfile('temp/'+base_name+'_stablized.npy') or args.re_preprocess:
+        imgs = []
+        for idx in tqdm(range(len(frames_pil)), desc="[Step 3] Stabilize the expression In Video:"):
+            if args.one_shot:
+                source_img = trans_image(frames_pil[0]).unsqueeze(0).to(device)
+                semantic_source_numpy = semantic_npy[0:1]
+            else:
+                source_img = trans_image(frames_pil[idx]).unsqueeze(0).to(device)
+                semantic_source_numpy = semantic_npy[idx:idx+1]
+            ratio = find_crop_norm_ratio(semantic_source_numpy, semantic_npy)
+            coeff = transform_semantic(semantic_npy, idx, ratio).unsqueeze(0).to(device)
+        
+            # hacking the new expression
+            coeff[:, :64, :] = expression[None, :64, None].to(device) 
+            with torch.no_grad():
+                output = D_Net(source_img, coeff)
+            img_stablized = np.uint8((output['fake_image'].squeeze(0).permute(1,2,0).cpu().clamp_(-1, 1).numpy() + 1 )/2. * 255)
+            imgs.append(cv2.cvtColor(img_stablized,cv2.COLOR_RGB2BGR)) 
+        np.save('temp/'+base_name+'_stablized.npy',imgs)
+        del D_Net
+    else:
+        print('[Step 3] Using saved stabilized video.')
+        imgs = np.load('temp/'+base_name+'_stablized.npy')
+    torch.cuda.empty_cache()
+
+    if not args.audio.endswith('.wav'):
+        command = 'ffmpeg -loglevel error -y -i {} -strict -2 {}'.format(args.audio, 'temp/{}/temp.wav'.format(args.tmp_dir))
+        subprocess.call(command, shell=True)
+        args.audio = 'temp/{}/temp.wav'.format(args.tmp_dir)
+    wav = audio.load_wav(args.audio, 16000)
+    mel = audio.melspectrogram(wav)
+    if np.isnan(mel.reshape(-1)).sum() > 0:
+        raise ValueError('Mel contains nan! Using a TTS voice? Add a small epsilon noise to the wav file and try again')
+
+    mel_step_size, mel_idx_multiplier, i, mel_chunks = 16, 80./fps, 0, []
+    while True:
+        start_idx = int(i * mel_idx_multiplier)
+        if start_idx + mel_step_size > len(mel[0]):
+            mel_chunks.append(mel[:, len(mel[0]) - mel_step_size:])
+            break
+        mel_chunks.append(mel[:, start_idx : start_idx + mel_step_size])
+        i += 1
+
+    print("[Step 4] Load audio; Length of mel chunks: {}".format(len(mel_chunks)))
+    imgs = imgs[:len(mel_chunks)]
+    full_frames = full_frames[:len(mel_chunks)]  
+    lm = lm[:len(mel_chunks)]
+    
+    imgs_enhanced = []
+    for idx in tqdm(range(len(imgs)), desc='[Step 5] Reference Enhancement'):
+        img = imgs[idx]
+        pred, _, _ = enhancer.process(img, img, face_enhance=True, possion_blending=False)
+        imgs_enhanced.append(pred)
+    gen = datagen(imgs_enhanced.copy(), mel_chunks, full_frames, None, (oy1,oy2,ox1,ox2))
+
+    frame_h, frame_w = full_frames[0].shape[:-1]
+    out = cv2.VideoWriter('temp/{}/result.mp4'.format(args.tmp_dir), cv2.VideoWriter_fourcc(*'mp4v'), fps, (frame_w, frame_h))
+    
+    if args.up_face != 'original':
+        instance = GANimationModel()
+        instance.initialize()
+        instance.setup()
+
+    kp_extractor = KeypointExtractor()
+    for i, (img_batch, mel_batch, frames, coords, img_original, f_frames) in enumerate(tqdm(gen, desc='[Step 6] Lip Synthesis:', total=int(np.ceil(float(len(mel_chunks)) / args.LNet_batch_size)))):
+        img_batch = torch.FloatTensor(np.transpose(img_batch, (0, 3, 1, 2))).to(device)
+        mel_batch = torch.FloatTensor(np.transpose(mel_batch, (0, 3, 1, 2))).to(device)
+        img_original = torch.FloatTensor(np.transpose(img_original, (0, 3, 1, 2))).to(device)/255. # BGR -> RGB
+        
+        with torch.no_grad():
+            incomplete, reference = torch.split(img_batch, 3, dim=1) 
+            pred, low_res = model(mel_batch, img_batch, reference)
+            pred = torch.clamp(pred, 0, 1)
+
+            if args.up_face in ['sad', 'angry', 'surprise']:
+                tar_aus = exp_aus_dict[args.up_face]
+            else:
+                pass
+            
+            if args.up_face == 'original':
+                cur_gen_faces = img_original
+            else:
+                test_batch = {'src_img': torch.nn.functional.interpolate((img_original * 2 - 1), size=(128, 128), mode='bilinear'), 
+                              'tar_aus': tar_aus.repeat(len(incomplete), 1)}
+                instance.feed_batch(test_batch)
+                instance.forward()
+                cur_gen_faces = torch.nn.functional.interpolate(instance.fake_img / 2. + 0.5, size=(384, 384), mode='bilinear')
+                
+            if args.without_rl1 is not False:
+                incomplete, reference = torch.split(img_batch, 3, dim=1)
+                mask = torch.where(incomplete==0, torch.ones_like(incomplete), torch.zeros_like(incomplete)) 
+                pred = pred * mask + cur_gen_faces * (1 - mask) 
+        
+        pred = pred.cpu().numpy().transpose(0, 2, 3, 1) * 255.
+
+        torch.cuda.empty_cache()
+        for p, f, xf, c in zip(pred, frames, f_frames, coords):
+            y1, y2, x1, x2 = c
+            p = cv2.resize(p.astype(np.uint8), (x2 - x1, y2 - y1))
+            
+            ff = xf.copy() 
+            ff[y1:y2, x1:x2] = p
+            
+            # month region enhancement by GFPGAN
+            cropped_faces, restored_faces, restored_img = restorer.enhance(
+                ff, has_aligned=False, only_center_face=True, paste_back=True)
+                # 0,   1,   2,   3,   4,   5,   6,   7,   8,  9, 10,  11,  12,
+            mm = [0,   0,   0,   0,   0,   0,   0,   0,   0,  0, 255, 255, 255, 0, 0, 0, 0, 0, 0]
+            mouse_mask = np.zeros_like(restored_img)
+            tmp_mask = enhancer.faceparser.process(restored_img[y1:y2, x1:x2], mm)[0]
+            mouse_mask[y1:y2, x1:x2]= cv2.resize(tmp_mask, (x2 - x1, y2 - y1))[:, :, np.newaxis] / 255.
+
+            height, width = ff.shape[:2]
+            restored_img, ff, full_mask = [cv2.resize(x, (512, 512)) for x in (restored_img, ff, np.float32(mouse_mask))]
+            img = Laplacian_Pyramid_Blending_with_mask(restored_img, ff, full_mask[:, :, 0], 10)
+            pp = np.uint8(cv2.resize(np.clip(img, 0 ,255), (width, height)))
+
+            pp, orig_faces, enhanced_faces = enhancer.process(pp, xf, bbox=c, face_enhance=False, possion_blending=True)
+            out.write(pp)
+    out.release()
+    
+    if not os.path.isdir(os.path.dirname(args.outfile)):
+        os.makedirs(os.path.dirname(args.outfile), exist_ok=True)
+    command = 'ffmpeg -loglevel error -y -i {} -i {} -strict -2 -q:v 1 {}'.format(args.audio, 'temp/{}/result.mp4'.format(args.tmp_dir), args.outfile)
+    subprocess.call(command, shell=platform.system() != 'Windows')
+    print('outfile:', args.outfile)
+
+
+# frames:256x256, full_frames: original size
+def datagen(frames, mels, full_frames, frames_pil, cox):
+    img_batch, mel_batch, frame_batch, coords_batch, ref_batch, full_frame_batch = [], [], [], [], [], []
+    base_name = args.face.split('/')[-1]
+    refs = []
+    image_size = 256 
+
+    # original frames
+    kp_extractor = KeypointExtractor()
+    fr_pil = [Image.fromarray(frame) for frame in frames]
+    lms = kp_extractor.extract_keypoint(fr_pil, 'temp/'+base_name+'x12_landmarks.txt')
+    frames_pil = [ (lm, frame) for frame,lm in zip(fr_pil, lms)] # frames is the croped version of modified face
+    crops, orig_images, quads  = crop_faces(image_size, frames_pil, scale=1.0, use_fa=True)
+    inverse_transforms = [calc_alignment_coefficients(quad + 0.5, [[0, 0], [0, image_size], [image_size, image_size], [image_size, 0]]) for quad in quads]
+    del kp_extractor.detector
+
+    oy1,oy2,ox1,ox2 = cox
+    face_det_results = face_detect(full_frames, args, jaw_correction=True)
+
+    for inverse_transform, crop, full_frame, face_det in zip(inverse_transforms, crops, full_frames, face_det_results):
+        imc_pil = paste_image(inverse_transform, crop, Image.fromarray(
+            cv2.resize(full_frame[int(oy1):int(oy2), int(ox1):int(ox2)], (256, 256))))
+
+        ff = full_frame.copy()
+        ff[int(oy1):int(oy2), int(ox1):int(ox2)] = cv2.resize(np.array(imc_pil.convert('RGB')), (ox2 - ox1, oy2 - oy1))
+        oface, coords = face_det
+        y1, y2, x1, x2 = coords
+        refs.append(ff[y1: y2, x1:x2])
+
+    for i, m in enumerate(mels):
+        idx = 0 if args.static else i % len(frames)
+        frame_to_save = frames[idx].copy()
+        face = refs[idx]
+        oface, coords = face_det_results[idx].copy()
+
+        face = cv2.resize(face, (args.img_size, args.img_size))
+        oface = cv2.resize(oface, (args.img_size, args.img_size))
+
+        img_batch.append(oface)
+        ref_batch.append(face) 
+        mel_batch.append(m)
+        coords_batch.append(coords)
+        frame_batch.append(frame_to_save)
+        full_frame_batch.append(full_frames[idx].copy())
+
+        if len(img_batch) >= args.LNet_batch_size:
+            img_batch, mel_batch, ref_batch = np.asarray(img_batch), np.asarray(mel_batch), np.asarray(ref_batch)
+            img_masked = img_batch.copy()
+            img_original = img_batch.copy()
+            img_masked[:, args.img_size//2:] = 0
+            img_batch = np.concatenate((img_masked, ref_batch), axis=3) / 255.
+            mel_batch = np.reshape(mel_batch, [len(mel_batch), mel_batch.shape[1], mel_batch.shape[2], 1])
+
+            yield img_batch, mel_batch, frame_batch, coords_batch, img_original, full_frame_batch
+            img_batch, mel_batch, frame_batch, coords_batch, img_original, full_frame_batch, ref_batch  = [], [], [], [], [], [], []
+
+    if len(img_batch) > 0:
+        img_batch, mel_batch, ref_batch = np.asarray(img_batch), np.asarray(mel_batch), np.asarray(ref_batch)
+        img_masked = img_batch.copy()
+        img_original = img_batch.copy()
+        img_masked[:, args.img_size//2:] = 0
+        img_batch = np.concatenate((img_masked, ref_batch), axis=3) / 255.
+        mel_batch = np.reshape(mel_batch, [len(mel_batch), mel_batch.shape[1], mel_batch.shape[2], 1])
+        yield img_batch, mel_batch, frame_batch, coords_batch, img_original, full_frame_batch
+
+
+if __name__ == '__main__':
+    main()

predict.py

@@ -0,0 +1,552 @@
+# Prediction interface for Cog ⚙️
+# https://github.com/replicate/cog/blob/main/docs/python.md
+
+import os
+import sys
+import argparse
+import subprocess
+import numpy as np
+from tqdm import tqdm
+from PIL import Image
+from scipy.io import loadmat
+import torch
+import cv2
+from cog import BasePredictor, Input, Path
+
+sys.path.insert(0, "third_part")
+sys.path.insert(0, "third_part/GPEN")
+sys.path.insert(0, "third_part/GFPGAN")
+
+# 3dmm extraction
+from third_part.face3d.util.preprocess import align_img
+from third_part.face3d.util.load_mats import load_lm3d
+from third_part.face3d.extract_kp_videos import KeypointExtractor
+
+# face enhancement
+from third_part.GPEN.gpen_face_enhancer import FaceEnhancement
+from third_part.GFPGAN.gfpgan import GFPGANer
+
+# expression control
+from third_part.ganimation_replicate.model.ganimation import GANimationModel
+
+from utils import audio
+from utils.ffhq_preprocess import Croper
+from utils.alignment_stit import crop_faces, calc_alignment_coefficients, paste_image
+from utils.inference_utils import (
+    Laplacian_Pyramid_Blending_with_mask,
+    face_detect,
+    load_model,
+    options,
+    split_coeff,
+    trans_image,
+    transform_semantic,
+    find_crop_norm_ratio,
+    load_face3d_net,
+    exp_aus_dict,
+)
+
+
+class Predictor(BasePredictor):
+    def setup(self) -> None:
+        """Load the model into memory to make running multiple predictions efficient"""
+        self.enhancer = FaceEnhancement(
+            base_dir="checkpoints",
+            size=512,
+            model="GPEN-BFR-512",
+            use_sr=False,
+            sr_model="rrdb_realesrnet_psnr",
+            channel_multiplier=2,
+            narrow=1,
+            device="cuda",
+        )
+        self.restorer = GFPGANer(
+            model_path="checkpoints/GFPGANv1.3.pth",
+            upscale=1,
+            arch="clean",
+            channel_multiplier=2,
+            bg_upsampler=None,
+        )
+        self.croper = Croper("checkpoints/shape_predictor_68_face_landmarks.dat")
+        self.kp_extractor = KeypointExtractor()
+
+        face3d_net_path = "checkpoints/face3d_pretrain_epoch_20.pth"
+
+        self.net_recon = load_face3d_net(face3d_net_path, "cuda")
+        self.lm3d_std = load_lm3d("checkpoints/BFM")
+
+    def predict(
+        self,
+        face: Path = Input(description="Input video file of a talking-head."),
+        input_audio: Path = Input(description="Input audio file."),
+    ) -> Path:
+        """Run a single prediction on the model"""
+        device = "cuda"
+        args = argparse.Namespace(
+            DNet_path="checkpoints/DNet.pt",
+            LNet_path="checkpoints/LNet.pth",
+            ENet_path="checkpoints/ENet.pth",
+            face3d_net_path="checkpoints/face3d_pretrain_epoch_20.pth",
+            face=str(face),
+            audio=str(input_audio),
+            exp_img="neutral",
+            outfile=None,
+            fps=25,
+            pads=[0, 20, 0, 0],
+            face_det_batch_size=4,
+            LNet_batch_size=16,
+            img_size=384,
+            crop=[0, -1, 0, -1],
+            box=[-1, -1, -1, -1],
+            nosmooth=False,
+            static=False,
+            up_face="original",
+            one_shot=False,
+            without_rl1=False,
+            tmp_dir="temp",
+            re_preprocess=False,
+        )
+
+        base_name = args.face.split("/")[-1]
+
+        if args.face.split(".")[1] in ["jpg", "png", "jpeg"]:
+            full_frames = [cv2.imread(args.face)]
+            args.static = True
+            fps = args.fps
+        else:
+            video_stream = cv2.VideoCapture(args.face)
+            fps = video_stream.get(cv2.CAP_PROP_FPS)
+            full_frames = []
+            while True:
+                still_reading, frame = video_stream.read()
+                if not still_reading:
+                    video_stream.release()
+                    break
+                y1, y2, x1, x2 = args.crop
+                if x2 == -1:
+                    x2 = frame.shape[1]
+                if y2 == -1:
+                    y2 = frame.shape[0]
+                frame = frame[y1:y2, x1:x2]
+                full_frames.append(frame)
+
+        full_frames_RGB = [
+            cv2.cvtColor(frame, cv2.COLOR_BGR2RGB) for frame in full_frames
+        ]
+        full_frames_RGB, crop, quad = self.croper.crop(full_frames_RGB, xsize=512)
+
+        clx, cly, crx, cry = crop
+        lx, ly, rx, ry = quad
+        lx, ly, rx, ry = int(lx), int(ly), int(rx), int(ry)
+        oy1, oy2, ox1, ox2 = (
+            cly + ly,
+            min(cly + ry, full_frames[0].shape[0]),
+            clx + lx,
+            min(clx + rx, full_frames[0].shape[1]),
+        )
+        # original_size = (ox2 - ox1, oy2 - oy1)
+        frames_pil = [
+            Image.fromarray(cv2.resize(frame, (256, 256))) for frame in full_frames_RGB
+        ]
+
+        # get the landmark according to the detected face.
+        if (
+            not os.path.isfile("temp/" + base_name + "_landmarks.txt")
+            or args.re_preprocess
+        ):
+            print("[Step 1] Landmarks Extraction in Video.")
+            lm = self.kp_extractor.extract_keypoint(
+                frames_pil, "./temp/" + base_name + "_landmarks.txt"
+            )
+        else:
+            print("[Step 1] Using saved landmarks.")
+            lm = np.loadtxt("temp/" + base_name + "_landmarks.txt").astype(np.float32)
+            lm = lm.reshape([len(full_frames), -1, 2])
+
+        if (
+            not os.path.isfile("temp/" + base_name + "_coeffs.npy")
+            or args.exp_img is not None
+            or args.re_preprocess
+        ):
+            video_coeffs = []
+            for idx in tqdm(
+                range(len(frames_pil)), desc="[Step 2] 3DMM Extraction In Video:"
+            ):
+                frame = frames_pil[idx]
+                W, H = frame.size
+                lm_idx = lm[idx].reshape([-1, 2])
+                if np.mean(lm_idx) == -1:
+                    lm_idx = (self.lm3d_std[:, :2] + 1) / 2.0
+                    lm_idx = np.concatenate([lm_idx[:, :1] * W, lm_idx[:, 1:2] * H], 1)
+                else:
+                    lm_idx[:, -1] = H - 1 - lm_idx[:, -1]
+
+                trans_params, im_idx, lm_idx, _ = align_img(
+                    frame, lm_idx, self.lm3d_std
+                )
+                trans_params = np.array(
+                    [float(item) for item in np.hsplit(trans_params, 5)]
+                ).astype(np.float32)
+                im_idx_tensor = (
+                    torch.tensor(np.array(im_idx) / 255.0, dtype=torch.float32)
+                    .permute(2, 0, 1)
+                    .to(device)
+                    .unsqueeze(0)
+                )
+                with torch.no_grad():
+                    coeffs = split_coeff(self.net_recon(im_idx_tensor))
+
+                pred_coeff = {key: coeffs[key].cpu().numpy() for key in coeffs}
+                pred_coeff = np.concatenate(
+                    [
+                        pred_coeff["id"],
+                        pred_coeff["exp"],
+                        pred_coeff["tex"],
+                        pred_coeff["angle"],
+                        pred_coeff["gamma"],
+                        pred_coeff["trans"],
+                        trans_params[None],
+                    ],
+                    1,
+                )
+                video_coeffs.append(pred_coeff)
+            semantic_npy = np.array(video_coeffs)[:, 0]
+            np.save("temp/" + base_name + "_coeffs.npy", semantic_npy)
+        else:
+            print("[Step 2] Using saved coeffs.")
+            semantic_npy = np.load("temp/" + base_name + "_coeffs.npy").astype(
+                np.float32
+            )
+
+        # generate the 3dmm coeff from a single image
+        if args.exp_img == "smile":
+            expression = torch.tensor(
+                loadmat("checkpoints/expression.mat")["expression_mouth"]
+            )[0]
+        else:
+            print("using expression center")
+            expression = torch.tensor(
+                loadmat("checkpoints/expression.mat")["expression_center"]
+            )[0]
+
+        # load DNet, model(LNet and ENet)
+        D_Net, model = load_model(args, device)
+
+        if (
+            not os.path.isfile("temp/" + base_name + "_stablized.npy")
+            or args.re_preprocess
+        ):
+            imgs = []
+            for idx in tqdm(
+                range(len(frames_pil)),
+                desc="[Step 3] Stabilize the expression In Video:",
+            ):
+                if args.one_shot:
+                    source_img = trans_image(frames_pil[0]).unsqueeze(0).to(device)
+                    semantic_source_numpy = semantic_npy[0:1]
+                else:
+                    source_img = trans_image(frames_pil[idx]).unsqueeze(0).to(device)
+                    semantic_source_numpy = semantic_npy[idx : idx + 1]
+                ratio = find_crop_norm_ratio(semantic_source_numpy, semantic_npy)
+                coeff = (
+                    transform_semantic(semantic_npy, idx, ratio).unsqueeze(0).to(device)
+                )
+
+                # hacking the new expression
+                coeff[:, :64, :] = expression[None, :64, None].to(device)
+                with torch.no_grad():
+                    output = D_Net(source_img, coeff)
+                img_stablized = np.uint8(
+                    (
+                        output["fake_image"]
+                        .squeeze(0)
+                        .permute(1, 2, 0)
+                        .cpu()
+                        .clamp_(-1, 1)
+                        .numpy()
+                        + 1
+                    )
+                    / 2.0
+                    * 255
+                )
+                imgs.append(cv2.cvtColor(img_stablized, cv2.COLOR_RGB2BGR))
+            np.save("temp/" + base_name + "_stablized.npy", imgs)
+            del D_Net
+        else:
+            print("[Step 3] Using saved stabilized video.")
+            imgs = np.load("temp/" + base_name + "_stablized.npy")
+        torch.cuda.empty_cache()
+
+        if not args.audio.endswith(".wav"):
+            command = "ffmpeg -loglevel error -y -i {} -strict -2 {}".format(
+                args.audio, "temp/{}/temp.wav".format(args.tmp_dir)
+            )
+            subprocess.call(command, shell=True)
+            args.audio = "temp/{}/temp.wav".format(args.tmp_dir)
+        wav = audio.load_wav(args.audio, 16000)
+        mel = audio.melspectrogram(wav)
+        if np.isnan(mel.reshape(-1)).sum() > 0:
+            raise ValueError(
+                "Mel contains nan! Using a TTS voice? Add a small epsilon noise to the wav file and try again"
+            )
+
+        mel_step_size, mel_idx_multiplier, i, mel_chunks = 16, 80.0 / fps, 0, []
+        while True:
+            start_idx = int(i * mel_idx_multiplier)
+            if start_idx + mel_step_size > len(mel[0]):
+                mel_chunks.append(mel[:, len(mel[0]) - mel_step_size :])
+                break
+            mel_chunks.append(mel[:, start_idx : start_idx + mel_step_size])
+            i += 1
+
+        print("[Step 4] Load audio; Length of mel chunks: {}".format(len(mel_chunks)))
+        imgs = imgs[: len(mel_chunks)]
+        full_frames = full_frames[: len(mel_chunks)]
+        lm = lm[: len(mel_chunks)]
+
+        imgs_enhanced = []
+        for idx in tqdm(range(len(imgs)), desc="[Step 5] Reference Enhancement"):
+            img = imgs[idx]
+            pred, _, _ = self.enhancer.process(
+                img, img, face_enhance=True, possion_blending=False
+            )
+            imgs_enhanced.append(pred)
+        gen = datagen(
+            imgs_enhanced.copy(), mel_chunks, full_frames, args, (oy1, oy2, ox1, ox2)
+        )
+
+        frame_h, frame_w = full_frames[0].shape[:-1]
+        out = cv2.VideoWriter(
+            "temp/{}/result.mp4".format(args.tmp_dir),
+            cv2.VideoWriter_fourcc(*"mp4v"),
+            fps,
+            (frame_w, frame_h),
+        )
+
+        if args.up_face != "original":
+            instance = GANimationModel()
+            instance.initialize()
+            instance.setup()
+
+        # kp_extractor = KeypointExtractor()
+        for i, (
+            img_batch,
+            mel_batch,
+            frames,
+            coords,
+            img_original,
+            f_frames,
+        ) in enumerate(
+            tqdm(
+                gen,
+                desc="[Step 6] Lip Synthesis:",
+                total=int(np.ceil(float(len(mel_chunks)) / args.LNet_batch_size)),
+            )
+        ):
+            img_batch = torch.FloatTensor(np.transpose(img_batch, (0, 3, 1, 2))).to(
+                device
+            )
+            mel_batch = torch.FloatTensor(np.transpose(mel_batch, (0, 3, 1, 2))).to(
+                device
+            )
+            img_original = (
+                torch.FloatTensor(np.transpose(img_original, (0, 3, 1, 2))).to(device)
+                / 255.0
+            )  # BGR -> RGB
+
+            with torch.no_grad():
+                incomplete, reference = torch.split(img_batch, 3, dim=1)
+                pred, low_res = model(mel_batch, img_batch, reference)
+                pred = torch.clamp(pred, 0, 1)
+
+                if args.up_face in ["sad", "angry", "surprise"]:
+                    tar_aus = exp_aus_dict[args.up_face]
+                else:
+                    pass
+
+                if args.up_face == "original":
+                    cur_gen_faces = img_original
+                else:
+                    test_batch = {
+                        "src_img": torch.nn.functional.interpolate(
+                            (img_original * 2 - 1), size=(128, 128), mode="bilinear"
+                        ),
+                        "tar_aus": tar_aus.repeat(len(incomplete), 1),
+                    }
+                    instance.feed_batch(test_batch)
+                    instance.forward()
+                    cur_gen_faces = torch.nn.functional.interpolate(
+                        instance.fake_img / 2.0 + 0.5, size=(384, 384), mode="bilinear"
+                    )
+
+                if args.without_rl1 is not False:
+                    incomplete, reference = torch.split(img_batch, 3, dim=1)
+                    mask = torch.where(
+                        incomplete == 0,
+                        torch.ones_like(incomplete),
+                        torch.zeros_like(incomplete),
+                    )
+                    pred = pred * mask + cur_gen_faces * (1 - mask)
+
+            pred = pred.cpu().numpy().transpose(0, 2, 3, 1) * 255.0
+
+            torch.cuda.empty_cache()
+            for p, f, xf, c in zip(pred, frames, f_frames, coords):
+                y1, y2, x1, x2 = c
+                p = cv2.resize(p.astype(np.uint8), (x2 - x1, y2 - y1))
+
+                ff = xf.copy()
+                ff[y1:y2, x1:x2] = p
+
+                # month region enhancement by GFPGAN
+                cropped_faces, restored_faces, restored_img = self.restorer.enhance(
+                    ff, has_aligned=False, only_center_face=True, paste_back=True
+                )
+                # 0,   1,   2,   3,   4,   5,   6,   7,   8,  9, 10,  11,  12,
+                mm = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 255, 255, 255, 0, 0, 0, 0, 0, 0]
+                mouse_mask = np.zeros_like(restored_img)
+                tmp_mask = self.enhancer.faceparser.process(
+                    restored_img[y1:y2, x1:x2], mm
+                )[0]
+                mouse_mask[y1:y2, x1:x2] = (
+                    cv2.resize(tmp_mask, (x2 - x1, y2 - y1))[:, :, np.newaxis] / 255.0
+                )
+
+                height, width = ff.shape[:2]
+                restored_img, ff, full_mask = [
+                    cv2.resize(x, (512, 512))
+                    for x in (restored_img, ff, np.float32(mouse_mask))
+                ]
+                img = Laplacian_Pyramid_Blending_with_mask(
+                    restored_img, ff, full_mask[:, :, 0], 10
+                )
+                pp = np.uint8(cv2.resize(np.clip(img, 0, 255), (width, height)))
+
+                pp, orig_faces, enhanced_faces = self.enhancer.process(
+                    pp, xf, bbox=c, face_enhance=False, possion_blending=True
+                )
+                out.write(pp)
+        out.release()
+
+        output_file = "/tmp/output.mp4"
+        command = "ffmpeg -loglevel error -y -i {} -i {} -strict -2 -q:v 1 {}".format(
+            args.audio, "temp/{}/result.mp4".format(args.tmp_dir), output_file
+        )
+        subprocess.call(command, shell=True)
+
+        return Path(output_file)
+
+
+# frames:256x256, full_frames: original size
+def datagen(frames, mels, full_frames, args, cox):
+    img_batch, mel_batch, frame_batch, coords_batch, ref_batch, full_frame_batch = (
+        [],
+        [],
+        [],
+        [],
+        [],
+        [],
+    )
+    base_name = args.face.split("/")[-1]
+    refs = []
+    image_size = 256
+
+    # original frames
+    kp_extractor = KeypointExtractor()
+    fr_pil = [Image.fromarray(frame) for frame in frames]
+    lms = kp_extractor.extract_keypoint(
+        fr_pil, "temp/" + base_name + "x12_landmarks.txt"
+    )
+    frames_pil = [
+        (lm, frame) for frame, lm in zip(fr_pil, lms)
+    ]  # frames is the croped version of modified face
+    crops, orig_images, quads = crop_faces(
+        image_size, frames_pil, scale=1.0, use_fa=True
+    )
+    inverse_transforms = [
+        calc_alignment_coefficients(
+            quad + 0.5,
+            [[0, 0], [0, image_size], [image_size, image_size], [image_size, 0]],
+        )
+        for quad in quads
+    ]
+    del kp_extractor.detector
+
+    oy1, oy2, ox1, ox2 = cox
+    face_det_results = face_detect(full_frames, args, jaw_correction=True)
+
+    for inverse_transform, crop, full_frame, face_det in zip(
+        inverse_transforms, crops, full_frames, face_det_results
+    ):
+        imc_pil = paste_image(
+            inverse_transform,
+            crop,
+            Image.fromarray(
+                cv2.resize(
+                    full_frame[int(oy1) : int(oy2), int(ox1) : int(ox2)], (256, 256)
+                )
+            ),
+        )
+
+        ff = full_frame.copy()
+        ff[int(oy1) : int(oy2), int(ox1) : int(ox2)] = cv2.resize(
+            np.array(imc_pil.convert("RGB")), (ox2 - ox1, oy2 - oy1)
+        )
+        oface, coords = face_det
+        y1, y2, x1, x2 = coords
+        refs.append(ff[y1:y2, x1:x2])
+
+    for i, m in enumerate(mels):
+        idx = 0 if args.static else i % len(frames)
+        frame_to_save = frames[idx].copy()
+        face = refs[idx]
+        oface, coords = face_det_results[idx].copy()
+
+        face = cv2.resize(face, (args.img_size, args.img_size))
+        oface = cv2.resize(oface, (args.img_size, args.img_size))
+
+        img_batch.append(oface)
+        ref_batch.append(face)
+        mel_batch.append(m)
+        coords_batch.append(coords)
+        frame_batch.append(frame_to_save)
+        full_frame_batch.append(full_frames[idx].copy())
+
+        if len(img_batch) >= args.LNet_batch_size:
+            img_batch, mel_batch, ref_batch = (
+                np.asarray(img_batch),
+                np.asarray(mel_batch),
+                np.asarray(ref_batch),
+            )
+            img_masked = img_batch.copy()
+            img_original = img_batch.copy()
+            img_masked[:, args.img_size // 2 :] = 0
+            img_batch = np.concatenate((img_masked, ref_batch), axis=3) / 255.0
+            mel_batch = np.reshape(
+                mel_batch, [len(mel_batch), mel_batch.shape[1], mel_batch.shape[2], 1]
+            )
+
+            yield img_batch, mel_batch, frame_batch, coords_batch, img_original, full_frame_batch
+            (
+                img_batch,
+                mel_batch,
+                frame_batch,
+                coords_batch,
+                img_original,
+                full_frame_batch,
+                ref_batch,
+            ) = ([], [], [], [], [], [], [])
+
+    if len(img_batch) > 0:
+        img_batch, mel_batch, ref_batch = (
+            np.asarray(img_batch),
+            np.asarray(mel_batch),
+            np.asarray(ref_batch),
+        )
+        img_masked = img_batch.copy()
+        img_original = img_batch.copy()
+        img_masked[:, args.img_size // 2 :] = 0
+        img_batch = np.concatenate((img_masked, ref_batch), axis=3) / 255.0
+        mel_batch = np.reshape(
+            mel_batch, [len(mel_batch), mel_batch.shape[1], mel_batch.shape[2], 1]
+        )
+        yield img_batch, mel_batch, frame_batch, coords_batch, img_original, full_frame_batch

requirements.txt

@@ -0,0 +1,15 @@
+gradio
+torch
+torchvision
+basicsr
+kornia
+face-alignment
+ninja
+einops
+facexlib
+librosa
+dlib
+numpy
+fastapi
+uvicorn
+ninja

temp/1.mp4_coeffs.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f99747e3debfae74d7169f06fa016e98385e081aa9301071db342dec38818588
+size 359592

temp/1.mp4_stablized.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55b8bc8e44ea961ce4c84ec162a45a6f337dbc95fe6dcf6711d66a1b4421fa70
+size 67436672

temp/dropbox5.mp4_coeffs.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae925c8a5488ee06eeef23e40dc3b17e91090d00373d0e5e670233fb015e5331
+size 351208

temp/dropbox5.mp4_stablized.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1e5b20415ce5616868467f44b0e557296b86e0b07ac76fbfae4355672c61dd1
+size 65863808

temp/face.mp4_coeffs.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96e6fbc6c00626ac512d50086324625a80e38d7e73c99227815df2a7db985760
+size 631024

temp/face.mp4_stablized.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:280d25163a5b9f6cfa00da6c57b1b68588874d711a458b003b5a0162047e55bf
+size 118358144

temp/temp/temp.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa99fa7f98235d432295627485004a4d67922055f679520bad65843a24fba75e
+size 1367060

third_part/GFPGAN/LICENSE

@@ -0,0 +1,351 @@
+Tencent is pleased to support the open source community by making GFPGAN available.
+
+Copyright (C) 2021 THL A29 Limited, a Tencent company.  All rights reserved.
+
+GFPGAN is licensed under the Apache License Version 2.0 except for the third-party components listed below.
+
+
+Terms of the Apache License Version 2.0:
+---------------------------------------------
+Apache License
+
+Version 2.0, January 2004
+
+http://www.apache.org/licenses/
+
+TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+1. Definitions.
+
+“License” shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document.
+
+“Licensor” shall mean the copyright owner or entity authorized by the copyright owner that is granting the License.
+
+“Legal Entity” shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, “control” means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity.
+
+“You” (or “Your”) shall mean an individual or Legal Entity exercising permissions granted by this License.
+
+“Source” form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files.
+
+“Object” form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types.
+
+“Work” shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below).
+
+“Derivative Works” shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof.
+
+“Contribution” shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, “submitted” means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as “Not a Contribution.”
+
+“Contributor” shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work.
+
+2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form.
+
+3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed.
+
+4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions:
+
+You must give any other recipients of the Work or Derivative Works a copy of this License; and
+
+You must cause any modified files to carry prominent notices stating that You changed the files; and
+
+You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and
+
+If the Work includes a “NOTICE” text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License.
+
+You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License.
+
+5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions.
+
+6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file.
+
+7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License.
+
+8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages.
+
+9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability.
+
+END OF TERMS AND CONDITIONS
+
+
+
+Other  dependencies and licenses:
+
+
+Open Source Software licensed under the Apache 2.0 license and Other Licenses of the Third-Party Components therein:
+---------------------------------------------
+1. basicsr
+Copyright 2018-2020 BasicSR Authors
+
+
+This BasicSR project is released under the Apache 2.0 license.
+
+A copy of Apache 2.0 is included in this file.
+
+StyleGAN2
+The codes are modified from the repository stylegan2-pytorch. Many thanks to the author - Kim Seonghyeon 😊 for translating from the official TensorFlow codes to PyTorch ones. Here is the license of stylegan2-pytorch.
+The official repository is https://github.com/NVlabs/stylegan2, and here is the NVIDIA license.
+DFDNet
+The codes are largely modified from the repository DFDNet. Their license is Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
+
+Terms of the Nvidia License:
+---------------------------------------------
+
+1. Definitions
+
+"Licensor" means any person or entity that distributes its Work.
+
+"Software" means the original work of authorship made available under
+this License.
+
+"Work" means the Software and any additions to or derivative works of
+the Software that are made available under this License.
+
+"Nvidia Processors" means any central processing unit (CPU), graphics
+processing unit (GPU), field-programmable gate array (FPGA),
+application-specific integrated circuit (ASIC) or any combination
+thereof designed, made, sold, or provided by Nvidia or its affiliates.
+
+The terms "reproduce," "reproduction," "derivative works," and
+"distribution" have the meaning as provided under U.S. copyright law;
+provided, however, that for the purposes of this License, derivative
+works shall not include works that remain separable from, or merely
+link (or bind by name) to the interfaces of, the Work.
+
+Works, including the Software, are "made available" under this License
+by including in or with the Work either (a) a copyright notice
+referencing the applicability of this License to the Work, or (b) a
+copy of this License.
+
+2. License Grants
+
+    2.1 Copyright Grant. Subject to the terms and conditions of this
+    License, each Licensor grants to you a perpetual, worldwide,
+    non-exclusive, royalty-free, copyright license to reproduce,
+    prepare derivative works of, publicly display, publicly perform,
+    sublicense and distribute its Work and any resulting derivative
+    works in any form.
+
+3. Limitations
+
+    3.1 Redistribution. You may reproduce or distribute the Work only
+    if (a) you do so under this License, (b) you include a complete
+    copy of this License with your distribution, and (c) you retain
+    without modification any copyright, patent, trademark, or
+    attribution notices that are present in the Work.
+
+    3.2 Derivative Works. You may specify that additional or different
+    terms apply to the use, reproduction, and distribution of your
+    derivative works of the Work ("Your Terms") only if (a) Your Terms
+    provide that the use limitation in Section 3.3 applies to your
+    derivative works, and (b) you identify the specific derivative
+    works that are subject to Your Terms. Notwithstanding Your Terms,
+    this License (including the redistribution requirements in Section
+    3.1) will continue to apply to the Work itself.
+
+    3.3 Use Limitation. The Work and any derivative works thereof only
+    may be used or intended for use non-commercially. The Work or
+    derivative works thereof may be used or intended for use by Nvidia
+    or its affiliates commercially or non-commercially. As used herein,
+    "non-commercially" means for research or evaluation purposes only.
+
+    3.4 Patent Claims. If you bring or threaten to bring a patent claim
+    against any Licensor (including any claim, cross-claim or
+    counterclaim in a lawsuit) to enforce any patents that you allege
+    are infringed by any Work, then your rights under this License from
+    such Licensor (including the grants in Sections 2.1 and 2.2) will
+    terminate immediately.
+
+    3.5 Trademarks. This License does not grant any rights to use any
+    Licensor's or its affiliates' names, logos, or trademarks, except
+    as necessary to reproduce the notices described in this License.
+
+    3.6 Termination. If you violate any term of this License, then your
+    rights under this License (including the grants in Sections 2.1 and
+    2.2) will terminate immediately.
+
+4. Disclaimer of Warranty.
+
+THE WORK IS PROVIDED "AS IS" WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WARRANTIES OR CONDITIONS OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE OR
+NON-INFRINGEMENT. YOU BEAR THE RISK OF UNDERTAKING ANY ACTIVITIES UNDER
+THIS LICENSE.
+
+5. Limitation of Liability.
+
+EXCEPT AS PROHIBITED BY APPLICABLE LAW, IN NO EVENT AND UNDER NO LEGAL
+THEORY, WHETHER IN TORT (INCLUDING NEGLIGENCE), CONTRACT, OR OTHERWISE
+SHALL ANY LICENSOR BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY DIRECT,
+INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF
+OR RELATED TO THIS LICENSE, THE USE OR INABILITY TO USE THE WORK
+(INCLUDING BUT NOT LIMITED TO LOSS OF GOODWILL, BUSINESS INTERRUPTION,
+LOST PROFITS OR DATA, COMPUTER FAILURE OR MALFUNCTION, OR ANY OTHER
+COMMERCIAL DAMAGES OR LOSSES), EVEN IF THE LICENSOR HAS BEEN ADVISED OF
+THE POSSIBILITY OF SUCH DAMAGES.
+
+MIT License
+
+Copyright (c) 2019 Kim Seonghyeon
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+
+
+Open Source Software licensed under the BSD 3-Clause license:
+---------------------------------------------
+1. torchvision
+Copyright (c) Soumith Chintala 2016,
+All rights reserved.
+
+2. torch
+Copyright (c) 2016-     Facebook, Inc            (Adam Paszke)
+Copyright (c) 2014-     Facebook, Inc            (Soumith Chintala)
+Copyright (c) 2011-2014 Idiap Research Institute (Ronan Collobert)
+Copyright (c) 2012-2014 Deepmind Technologies    (Koray Kavukcuoglu)
+Copyright (c) 2011-2012 NEC Laboratories America (Koray Kavukcuoglu)
+Copyright (c) 2011-2013 NYU                      (Clement Farabet)
+Copyright (c) 2006-2010 NEC Laboratories America (Ronan Collobert, Leon Bottou, Iain Melvin, Jason Weston)
+Copyright (c) 2006      Idiap Research Institute (Samy Bengio)
+Copyright (c) 2001-2004 Idiap Research Institute (Ronan Collobert, Samy Bengio, Johnny Mariethoz)
+
+
+Terms of the BSD 3-Clause License:
+---------------------------------------------
+Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
+
+3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+
+
+Open Source Software licensed under the BSD 3-Clause License and Other Licenses of the Third-Party Components therein:
+---------------------------------------------
+1. numpy
+Copyright (c) 2005-2020, NumPy Developers.
+All rights reserved.
+
+A copy of BSD 3-Clause License is included in this file.
+
+The NumPy repository and source distributions bundle several libraries that are
+compatibly licensed.  We list these here.
+
+Name: Numpydoc
+Files: doc/sphinxext/numpydoc/*
+License: BSD-2-Clause
+  For details, see doc/sphinxext/LICENSE.txt
+
+Name: scipy-sphinx-theme
+Files: doc/scipy-sphinx-theme/*
+License: BSD-3-Clause AND PSF-2.0 AND Apache-2.0
+  For details, see doc/scipy-sphinx-theme/LICENSE.txt
+
+Name: lapack-lite
+Files: numpy/linalg/lapack_lite/*
+License: BSD-3-Clause
+  For details, see numpy/linalg/lapack_lite/LICENSE.txt
+
+Name: tempita
+Files: tools/npy_tempita/*
+License: MIT
+  For details, see tools/npy_tempita/license.txt
+
+Name: dragon4
+Files: numpy/core/src/multiarray/dragon4.c
+License: MIT
+  For license text, see numpy/core/src/multiarray/dragon4.c
+
+
+
+Open Source Software licensed under the MIT license:
+---------------------------------------------
+1. facexlib
+Copyright (c) 2020 Xintao Wang
+
+2. opencv-python
+Copyright (c) Olli-Pekka Heinisuo
+Please note that only files in cv2 package are used.
+
+
+Terms of the MIT License:
+---------------------------------------------
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+
+
+Open Source Software licensed under the MIT license and Other Licenses of the Third-Party Components therein:
+---------------------------------------------
+1. tqdm
+Copyright (c) 2013 noamraph
+
+`tqdm` is a product of collaborative work.
+Unless otherwise stated, all authors (see commit logs) retain copyright
+for their respective work, and release the work under the MIT licence
+(text below).
+
+Exceptions or notable authors are listed below
+in reverse chronological order:
+
+* files: *
+  MPLv2.0 2015-2020 (c) Casper da Costa-Luis
+  [casperdcl](https://github.com/casperdcl).
+* files: tqdm/_tqdm.py
+  MIT 2016 (c) [PR #96] on behalf of Google Inc.
+* files: tqdm/_tqdm.py setup.py README.rst MANIFEST.in .gitignore
+  MIT 2013 (c) Noam Yorav-Raphael, original author.
+
+[PR #96]: https://github.com/tqdm/tqdm/pull/96
+
+
+Mozilla Public Licence (MPL) v. 2.0 - Exhibit A
+-----------------------------------------------
+
+This Source Code Form is subject to the terms of the
+Mozilla Public License, v. 2.0.
+If a copy of the MPL was not distributed with this file,
+You can obtain one at https://mozilla.org/MPL/2.0/.
+
+
+MIT License (MIT)
+-----------------
+
+Copyright (c) 2013 noamraph
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
+the Software, and to permit persons to whom the Software is furnished to do so,
+subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
+FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

third_part/GFPGAN/gfpgan/__init__.py

@@ -0,0 +1,8 @@
+# flake8: noqa
+
+from .archs import *
+from .data import *
+from .models import *
+from .utils import *
+
+# from .version import *

third_part/GFPGAN/gfpgan/archs/__init__.py

@@ -0,0 +1,10 @@
+import importlib
+from basicsr.utils import scandir
+from os import path as osp
+
+# automatically scan and import arch modules for registry
+# scan all the files that end with '_arch.py' under the archs folder
+arch_folder = osp.dirname(osp.abspath(__file__))
+arch_filenames = [osp.splitext(osp.basename(v))[0] for v in scandir(arch_folder) if v.endswith('_arch.py')]
+# import all the arch modules
+_arch_modules = [importlib.import_module(f'gfpgan.archs.{file_name}') for file_name in arch_filenames]