mirror of
https://github.com/iperov/DeepFaceLab.git
synced 2025-03-12 20:42:45 -07:00
45582d129d
with XSeg model you can train your own mask segmentator of dst(and src) faces that will be used in merger for whole_face. Instead of using a pretrained model (which does not exist), you control which part of faces should be masked. Workflow is not easy, but at the moment it is the best solution for obtaining the best quality of whole_face's deepfakes using minimum effort without rotoscoping in AfterEffects. new scripts: XSeg) data_dst edit.bat XSeg) data_dst merge.bat XSeg) data_dst split.bat XSeg) data_src edit.bat XSeg) data_src merge.bat XSeg) data_src split.bat XSeg) train.bat Usage: unpack dst faceset if packed run XSeg) data_dst split.bat this scripts extracts (previously saved) .json data from jpg faces to use in label tool. run XSeg) data_dst edit.bat new tool 'labelme' is used use polygon (CTRL-N) to mask the face name polygon "1" (one symbol) as include polygon name polygon "0" (one symbol) as exclude polygon 'exclude polygons' will be applied after all 'include polygons' Hot keys: ctrl-N create polygon ctrl-J edit polygon A/D navigate between frames ctrl + mousewheel image zoom mousewheel vertical scroll alt+mousewheel horizontal scroll repeat for 10/50/100 faces, you don't need to mask every frame of dst, only frames where the face is different significantly, for example: closed eyes changed head direction changed light the more various faces you mask, the more quality you will get Start masking from the upper left area and follow the clockwise direction. Keep the same logic of masking for all frames, for example: the same approximated jaw line of the side faces, where the jaw is not visible the same hair line Mask the obstructions using polygon with name "0". run XSeg) data_dst merge.bat this script merges .json data of polygons into jpg faces, therefore faceset can be sorted or packed as usual. run XSeg) train.bat train the model Check the faces of 'XSeg dst faces' preview. if some faces have wrong or glitchy mask, then repeat steps: split run edit find these glitchy faces and mask them merge train further or restart training from scratch Restart training of XSeg model is only possible by deleting all 'model\XSeg_*' files. If you want to get the mask of the predicted face in merger, you should repeat the same steps for src faceset. New mask modes available in merger for whole_face: XSeg-prd - XSeg mask of predicted face -> faces from src faceset should be labeled XSeg-dst - XSeg mask of dst face -> faces from dst faceset should be labeled XSeg-prd*XSeg-dst - the smallest area of both if workspace\model folder contains trained XSeg model, then merger will use it, otherwise you will get transparent mask by using XSeg-* modes. Some screenshots: label tool: https://i.imgur.com/aY6QGw1.jpg trainer : https://i.imgur.com/NM1Kn3s.jpg merger : https://i.imgur.com/glUzFQ8.jpg example of the fake using 13 segmented dst faces : https://i.imgur.com/wmvyizU.gifv
91 lines
3.0 KiB
Python
91 lines
3.0 KiB
Python
import os
|
|
import pickle
|
|
from functools import partial
|
|
from pathlib import Path
|
|
|
|
import cv2
|
|
import numpy as np
|
|
|
|
from core.interact import interact as io
|
|
from core.leras import nn
|
|
|
|
|
|
class XSegNet(object):
|
|
VERSION = 1
|
|
|
|
def __init__ (self, name,
|
|
resolution,
|
|
load_weights=True,
|
|
weights_file_root=None,
|
|
training=False,
|
|
place_model_on_cpu=False,
|
|
run_on_cpu=False,
|
|
optimizer=None,
|
|
data_format="NHWC"):
|
|
|
|
nn.initialize(data_format=data_format)
|
|
tf = nn.tf
|
|
|
|
self.weights_file_root = Path(weights_file_root) if weights_file_root is not None else Path(__file__).parent
|
|
|
|
with tf.device ('/CPU:0'):
|
|
#Place holders on CPU
|
|
self.input_t = tf.placeholder (nn.floatx, nn.get4Dshape(resolution,resolution,3) )
|
|
self.target_t = tf.placeholder (nn.floatx, nn.get4Dshape(resolution,resolution,1) )
|
|
|
|
# Initializing model classes
|
|
with tf.device ('/CPU:0' if place_model_on_cpu else '/GPU:0'):
|
|
self.model = nn.XSeg(3, 32, 1, name=name)
|
|
self.model_weights = self.model.get_weights()
|
|
|
|
model_name = f'{name}_{resolution}'
|
|
|
|
self.model_filename_list = [ [self.model, f'{model_name}.npy'] ]
|
|
|
|
if training:
|
|
if optimizer is None:
|
|
raise ValueError("Optimizer should be provided for training mode.")
|
|
|
|
self.opt = optimizer
|
|
self.opt.initialize_variables (self.model_weights, vars_on_cpu=place_model_on_cpu)
|
|
self.model_filename_list += [ [self.opt, f'{model_name}_opt.npy' ] ]
|
|
else:
|
|
with tf.device ('/CPU:0' if run_on_cpu else '/GPU:0'):
|
|
_, pred = self.model(self.input_t)
|
|
|
|
def net_run(input_np):
|
|
return nn.tf_sess.run ( [pred], feed_dict={self.input_t :input_np})[0]
|
|
self.net_run = net_run
|
|
|
|
# Loading/initializing all models/optimizers weights
|
|
for model, filename in self.model_filename_list:
|
|
do_init = not load_weights
|
|
|
|
if not do_init:
|
|
do_init = not model.load_weights( self.weights_file_root / filename )
|
|
|
|
if do_init:
|
|
model.init_weights()
|
|
|
|
def flow(self, x):
|
|
return self.model(x)
|
|
|
|
def get_weights(self):
|
|
return self.model_weights
|
|
|
|
def save_weights(self):
|
|
for model, filename in io.progress_bar_generator(self.model_filename_list, "Saving", leave=False):
|
|
model.save_weights( self.weights_file_root / filename )
|
|
|
|
def extract (self, input_image):
|
|
input_shape_len = len(input_image.shape)
|
|
if input_shape_len == 3:
|
|
input_image = input_image[None,...]
|
|
|
|
result = np.clip ( self.net_run(input_image), 0, 1.0 )
|
|
result[result < 0.1] = 0 #get rid of noise
|
|
|
|
if input_shape_len == 3:
|
|
result = result[0]
|
|
|
|
return result |