File size: 9,180 Bytes
4e42a1b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 |
from glob import glob
import shutil
import torch
from time import strftime
import os, sys, time
from argparse import ArgumentParser
import platform
from src.utils.preprocess import CropAndExtract
from src.test_audio2coeff import Audio2Coeff
from src.facerender.animate import AnimateFromCoeff
from src.facerender.pirender_animate import AnimateFromCoeff_PIRender
from src.generate_batch import get_data
from src.generate_facerender_batch import get_facerender_data
from src.utils.init_path import init_path
import random
import numpy as np
def set_seed(seed):
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
if torch.cuda.is_available():
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
torch.backends.cudnn.deterministic = True
def main(args):
#torch.backends.cudnn.enabled = False
set_seed(42)
# args.facerender = 'pirender'
pic_path = args.source_image
audio_path = args.driven_audio
save_dir = args.result_dir
os.makedirs(save_dir, exist_ok=True)
pose_style = args.pose_style
device = args.device
batch_size = args.batch_size
input_yaw_list = args.input_yaw
input_pitch_list = args.input_pitch
input_roll_list = args.input_roll
ref_eyeblink = args.ref_eyeblink
ref_pose = args.ref_pose
# print(args.still)
# assert False
current_root_path = os.path.split(sys.argv[0])[0]
sadtalker_paths = init_path(args.checkpoint_dir, os.path.join(current_root_path, 'src/config'), args.size, args.old_version, args.preprocess)
#init model
preprocess_model = CropAndExtract(sadtalker_paths, device)
audio_to_coeff = Audio2Coeff(sadtalker_paths, device)
if args.facerender == 'facevid2vid':
animate_from_coeff = AnimateFromCoeff(sadtalker_paths, device)
elif args.facerender == 'pirender':
animate_from_coeff = AnimateFromCoeff_PIRender(sadtalker_paths, device)
else:
raise(RuntimeError('Unknown model: {}'.format(args.facerender)))
#crop image and extract 3dmm from image
first_frame_dir = os.path.join(save_dir, 'first_frame_dir')
os.makedirs(first_frame_dir, exist_ok=True)
print('3DMM Extraction for source image')
first_coeff_path, crop_pic_path, crop_info = preprocess_model.generate(pic_path, first_frame_dir, args.preprocess,\
source_image_flag=True, pic_size=args.size)
if first_coeff_path is None:
print("Can't get the coeffs of the input")
return
if ref_eyeblink is not None:
ref_eyeblink_videoname = os.path.splitext(os.path.split(ref_eyeblink)[-1])[0]
ref_eyeblink_frame_dir = os.path.join(save_dir, ref_eyeblink_videoname)
os.makedirs(ref_eyeblink_frame_dir, exist_ok=True)
print('3DMM Extraction for the reference video providing eye blinking')
ref_eyeblink_coeff_path, _, _ = preprocess_model.generate(ref_eyeblink, ref_eyeblink_frame_dir, args.preprocess, source_image_flag=False)
else:
ref_eyeblink_coeff_path=None
if ref_pose is not None:
if ref_pose == ref_eyeblink:
ref_pose_coeff_path = ref_eyeblink_coeff_path
else:
ref_pose_videoname = os.path.splitext(os.path.split(ref_pose)[-1])[0]
ref_pose_frame_dir = os.path.join(save_dir, ref_pose_videoname)
os.makedirs(ref_pose_frame_dir, exist_ok=True)
print('3DMM Extraction for the reference video providing pose')
ref_pose_coeff_path, _, _ = preprocess_model.generate(ref_pose, ref_pose_frame_dir, args.preprocess, source_image_flag=False)
else:
ref_pose_coeff_path=None
#audio2ceoff
batch = get_data(first_coeff_path, audio_path, device, ref_eyeblink_coeff_path, still=args.still)
coeff_path = audio_to_coeff.generate(batch, save_dir, pose_style, ref_pose_coeff_path)
# print(ref_pose_coeff_path.shape)
# print(coeff_path.shape)
# assert False
# 3dface render
if args.face3dvis:
from src.face3d.visualize import gen_composed_video
gen_composed_video(args, device, first_coeff_path, coeff_path, audio_path, \
os.path.join(save_dir, '3dface.mp4'), os.path.join(save_dir, 'landmarks.mp4'), crop_info, extended_crop= True if 'ext' in args.preprocess else False )
return
#coeff2video
data = get_facerender_data(coeff_path, crop_pic_path, first_coeff_path, audio_path,
batch_size, input_yaw_list, input_pitch_list, input_roll_list,
expression_scale=args.expression_scale, still_mode=args.still, preprocess=args.preprocess, size=args.size, facemodel=args.facerender)
result = animate_from_coeff.generate(data, save_dir, pic_path, crop_info, \
enhancer=args.enhancer, background_enhancer=args.background_enhancer, preprocess=args.preprocess, img_size=args.size)
shutil.move(result, save_dir+'.mp4')
print('The generated video is named:', save_dir+'.mp4')
# result = animate_from_coeff.generate_flow(data, args.result_dir, pic_path, crop_info, \
# enhancer=args.enhancer, background_enhancer=args.background_enhancer, preprocess=args.preprocess, img_size=args.size)
# if not args.verbose:
# shutil.rmtree(save_dir)
if __name__ == '__main__':
parser = ArgumentParser()
parser.add_argument("--driven_audio", default='./examples/driven_audio/bus_chinese.wav', help="path to driven audio")
parser.add_argument("--source_image", default='./examples/source_image/full_body_1.png', help="path to source image")
parser.add_argument("--ref_eyeblink", default=None, help="path to reference video providing eye blinking")
parser.add_argument("--ref_pose", default=None, help="path to reference video providing pose")
parser.add_argument("--checkpoint_dir", default='./ckpts/sad_talker', help="path to output")
parser.add_argument("--result_dir", default='./results', help="path to output")
parser.add_argument("--pose_style", type=int, default=0, help="input pose style from [0, 46)")
parser.add_argument("--batch_size", type=int, default=1, help="the batch size of facerender")
parser.add_argument("--size", type=int, default=256, help="the image size of the facerender")
parser.add_argument("--expression_scale", type=float, default=1., help="the batch size of facerender")
parser.add_argument('--input_yaw', nargs='+', type=int, default=None, help="the input yaw degree of the user ")
parser.add_argument('--input_pitch', nargs='+', type=int, default=None, help="the input pitch degree of the user")
parser.add_argument('--input_roll', nargs='+', type=int, default=None, help="the input roll degree of the user")
parser.add_argument('--enhancer', type=str, default=None, help="Face enhancer, [gfpgan, RestoreFormer]")
parser.add_argument('--background_enhancer', type=str, default=None, help="background enhancer, [realesrgan]")
parser.add_argument("--cpu", dest="cpu", action="store_true")
parser.add_argument("--face3dvis", action="store_true", help="generate 3d face and 3d landmarks")
parser.add_argument("--still", action="store_true", help="can crop back to the original videos for the full body aniamtion")
parser.add_argument("--preprocess", default='crop', choices=['crop', 'extcrop', 'resize', 'full', 'extfull'], help="how to preprocess the images" )
parser.add_argument("--verbose",action="store_true", help="saving the intermedia output or not" )
parser.add_argument("--old_version",action="store_true", help="use the pth other than safetensor version" )
parser.add_argument("--facerender", default='facevid2vid', choices=['pirender', 'facevid2vid'] )
# net structure and parameters
parser.add_argument('--net_recon', type=str, default='resnet50', choices=['resnet18', 'resnet34', 'resnet50'], help='useless')
parser.add_argument('--init_path', type=str, default=None, help='Useless')
parser.add_argument('--use_last_fc',default=False, help='zero initialize the last fc')
parser.add_argument('--bfm_folder', type=str, default='./ckpts/sad_talker/BFM_Fitting/')
parser.add_argument('--bfm_model', type=str, default='BFM_model_front.mat', help='bfm model')
# default renderer parameters
parser.add_argument('--focal', type=float, default=1015.)
parser.add_argument('--center', type=float, default=112.)
parser.add_argument('--camera_d', type=float, default=10.)
parser.add_argument('--z_near', type=float, default=5.)
parser.add_argument('--z_far', type=float, default=15.)
args = parser.parse_args()
if torch.cuda.is_available() and not args.cpu:
args.device = "cuda"
elif platform.system() == 'Darwin' and args.facerender == 'pirender': # macos
args.device = "mps"
else:
args.device = "cpu"
main(args)
|