Status:Executing -train.py
02/09/2019 09:33:38 INFO Log level set to: INFO
Using TensorFlow backend.
02/09/2019 09:33:39 INFO Model A Directory: /home/afda/faceswap/faceswap-master/output
02/09/2019 09:33:39 INFO Model B Directory: /home/afda/faceswap/faceswap-master/plugins/extract/align
02/09/2019 09:33:39 INFO Training data directory: /home/afda/faceswap/faceswap-master/models
02/09/2019 09:33:39 INFO Loading data, this may take a while...
02/09/2019 09:33:39 INFO Starting. Press 'ENTER' to stop training and save model
02/09/2019 09:33:39 INFO Loading Model from Model_Original plugin...
02/09/2019 09:33:39 WARNING Error loading training info: No such file or directory
02/09/2019 09:33:39 WARNING Failed loading existing training data. Starting a fresh model: /home/afda
/faceswap
/faceswap-master/models
02/09/2019 09:33:39 INFO Loading Trainer from Model_Original plugin...
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/home/afda/faceswap/faceswap-master/lib/multithreading.py", line 187, in run
for item in self.generator:
File "/home/afda/faceswap/faceswap-master/lib/training_data.py", line 23, in minibatch
assert length >= batchsize, "Number of images is lower than batch-size (Note that too few images may lead to bad
training). # images: {}, batch-size: {}".format(length, batchsize)
AssertionError: Number of images is lower than batch-size (Note that too few images may lead to bad training). #
images: 33, batch-size: 64
至少数据量为64张图片，而数据量只有33个，所以导致换脸失败，如果实在没有那么多图片可以复制一下来增加数据量欺骗tensorflow。。
换脸要求双方都有效才会效果好
而且换脸的顺序也有问题。。。input1和input2之间到底是怎么回事