Page 1 of 1

Faceswap Gui Aborts?

Posted: Mon Sep 07, 2020 10:16 pm
by DrDreGFunk

Hey guys, I'm hoping this will be a simple fix as I'm new to this -

When I try to start training it'll just say at the bottom "Process Exited" and "Status: Aborted - train.py". I don't know whether if it's because of a CPU problem or if I'm missing something? I also ran into the same "Process Exited" / "Abort" problem with the S3Fd detector when extracting - I know it says GPU only, but I swear I've read somewhere on here that CPU works, it's just very slow? Got the extraction working with Mtcnn anyway, but would use S3Fd in the future if it does work and fixed this abort situation. Oh, I don't know if this is a bug, but when I went to clean the Alignment files to remove the unwanted faces, it gave me an error (I can't remember exactly what it was, but I think it said something about "Traceback 1"?) However, when I ran it through again by reselecting the file again from the browser pop up window, it strangely worked. Not sure if it corrupted it or something though :?

Code: Select all

Loading...
Setting Faceswap backend to AMD
09/07/2020 22:20:26 INFO     Log level set to: TRACE
09/07/2020 22:20:26 INFO     Setting up for PlaidML
09/07/2020 22:20:26 VERBOSE  Setting Keras Backend to PlaidML
09/07/2020 22:20:26 INFO     Using GPU(s): ['Intel - Iris Pro (experimental)']
09/07/2020 22:20:26 INFO     Successfully set up for PlaidML
09/07/2020 22:20:29 INFO     Model A Directory: /Volumes/nameUSB/DF/faceswap-master/Kevin_faces
09/07/2020 22:20:29 INFO     Model B Directory: /Volumes/nameUSB/DF/faceswap-master/TheRock_faces
09/07/2020 22:20:29 INFO     Training data directory: /Volumes/nameUSB/DF/faceswap-master/model
09/07/2020 22:20:29 INFO     ===================================================
09/07/2020 22:20:29 INFO       Starting
09/07/2020 22:20:29 INFO       Press 'Stop' to save and quit
09/07/2020 22:20:29 INFO     ===================================================
09/07/2020 22:20:30 INFO     Loading data, this may take a while...
09/07/2020 22:20:30 INFO     Loading Model from Original plugin...
Using plaidml.keras.backend backend.
09/07/2020 22:20:30 VERBOSE  Loading config: '/Users/name/faceswap/config/train.ini'
09/07/2020 22:20:30 VERBOSE  Loading config: '/Users/name/faceswap/config/train.ini'
09/07/2020 22:20:30 INFO     No existing state file found. Generating.
09/07/2020 22:20:30 INFO     Opening device "opencl_intel_iris_pro.0"
09/07/2020 22:20:31 VERBOSE  Using Adam optimizer
09/07/2020 22:20:31 VERBOSE  __________________________________________________________________________________________________
09/07/2020 22:20:31 VERBOSE  Layer (type)                    Output Shape         Param #     Connected to
09/07/2020 22:20:31 VERBOSE  ==================================================================================================
09/07/2020 22:20:31 VERBOSE  face_in_a (InputLayer)          (None, 64, 64, 3)    0
09/07/2020 22:20:31 VERBOSE  __________________________________________________________________________________________________
09/07/2020 22:20:31 VERBOSE  face_in_b (InputLayer)          (None, 64, 64, 3)    0
09/07/2020 22:20:31 VERBOSE  __________________________________________________________________________________________________
09/07/2020 22:20:31 VERBOSE  encoder (Model)                 (None, 8, 8, 512)    69662976    face_in_a[0][0]
09/07/2020 22:20:31 VERBOSE                                                                   face_in_b[0][0]
09/07/2020 22:20:31 VERBOSE  __________________________________________________________________________________________________

09/07/2020 22:20:31 VERBOSE  decoder_a (Model)               (None, 64, 64, 3)    6199747     encoder[1][0]
09/07/2020 22:20:31 VERBOSE  __________________________________________________________________________________________________
09/07/2020 22:20:31 VERBOSE  decoder_b (Model)               (None, 64, 64, 3)    6199747     encoder[2][0]Reading training images (A):   0%|          | 0/1959 [00:00<?, ?it/s]

09/07/2020 22:20:31 VERBOSE  ==================================================================================================
09/07/2020 22:20:31 VERBOSE  Total params: 82,062,470
09/07/2020 22:20:31 VERBOSE  Trainable params: 82,062,470
09/07/2020 22:20:31 VERBOSE  Non-trainable params: 0
09/07/2020 22:20:31 VERBOSE  __________________________________________________________________________________________________
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  Layer (type)                 Output Shape              Param #
09/07/2020 22:20:31 VERBOSE  =================================================================Reading training images (A):   0%|          | 1/1959 [00:00<03:18,  9.89it/s]

09/07/2020 22:20:31 VERBOSE  input_1 (InputLayer)         (None, 64, 64, 3)         0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_128_0_conv2d (Conv2D)   (None, 32, 32, 128)       9728
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_128_0_leakyrelu (LeakyR (None, 32, 32, 128)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________Reading training images (A):   1%|▏         | 26/1959 [00:00<02:19, 13.87it/s]

09/07/2020 22:20:31 VERBOSE  conv_256_0_conv2d (Conv2D)   (None, 16, 16, 256)       819456
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_256_0_leakyrelu (LeakyR (None, 16, 16, 256)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_512_0_conv2d (Conv2D)   (None, 8, 8, 512)         3277312
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_512_0_leakyrelu (LeakyR (None, 8, 8, 512)         0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_1024_0_conv2d (Conv2D)  (None, 4, 4, 1024)        13108224Reading training images (A):   2%|▏         | 41/1959 [00:00<01:40, 19.03it/s]

09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  conv_1024_0_leakyrelu (Leaky (None, 4, 4, 1024)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  flatten_1 (Flatten)          (None, 16384)             0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  dense_1 (Dense)              (None, 1024)              16778240
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  dense_2 (Dense)              (None, 16384)             16793600
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  reshape_1 (Reshape)          (None, 4, 4, 1024)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_512_0_conv2d_conv2d  (None, 4, 4, 2048)        18876416
Reading training images (A):  14%|█▍        | 278/1959 [00:00<01:02, 27.09it/s]09/07/2020 22:20:31 VERBOSE  _________________________________________________________________

09/07/2020 22:20:31 VERBOSE  upscale_512_0_conv2d_leakyre (None, 4, 4, 2048)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_512_0_pixelshuffler  (None, 8, 8, 512)         0
09/07/2020 22:20:31 VERBOSE  =================================================================
09/07/2020 22:20:31 VERBOSE  Total params: 69,662,976
09/07/2020 22:20:31 VERBOSE  Trainable params: 69,662,976
09/07/2020 22:20:31 VERBOSE  Non-trainable params: 0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  Layer (type)                 Output Shape              Param #
09/07/2020 22:20:31 VERBOSE  =================================================================
09/07/2020 22:20:31 VERBOSE  input_2 (InputLayer)         (None, 8, 8, 512)         0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_256_0_conv2d_conv2d  (None, 8, 8, 1024)        4719616Reading training images (A):  29%|██▉       | 564/1959 [00:00<00:36, 38.41it/s]

09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_256_0_conv2d_leakyre (None, 8, 8, 1024)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_256_0_pixelshuffler  (None, 16, 16, 256)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_128_0_conv2d_conv2d  (None, 16, 16, 512)       1180160
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_128_0_conv2d_leakyre (None, 16, 16, 512)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_128_0_pixelshuffler  (None, 32, 32, 128)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_64_0_conv2d_conv2d ( (None, 32, 32, 256)       295168
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_64_0_conv2d_leakyrel (None, 32, 32, 256)       0Reading training images (A):  34%|███▍      | 665/1959 [00:00<00:24, 53.74it/s]

09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_64_0_pixelshuffler ( (None, 64, 64, 64)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  face_out_a_conv2d (Conv2D)   (None, 64, 64, 3)         4803
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  face_out_a (Activation)      (None, 64, 64, 3)         0
09/07/2020 22:20:31 VERBOSE  =================================================================
09/07/2020 22:20:31 VERBOSE  Total params: 6,199,747
09/07/2020 22:20:31 VERBOSE  Trainable params: 6,199,747
09/07/2020 22:20:31 VERBOSE  Non-trainable params: 0Reading training images (A):  39%|███▊      | 759/1959 [00:00<00:16, 73.51it/s]

09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  Layer (type)                 Output Shape              Param #
09/07/2020 22:20:31 VERBOSE  =================================================================
09/07/2020 22:20:31 VERBOSE  input_3 (InputLayer)         (None, 8, 8, 512)         0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_256_1_conv2d_conv2d  (None, 8, 8, 1024)        4719616
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_256_1_conv2d_leakyre (None, 8, 8, 1024)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
Reading training images (A):  55%|█████▍    | 1072/1959 [00:01<00:08, 103.29it/s]
09/07/2020 22:20:31 VERBOSE  upscale_256_1_pixelshuffler  (None, 16, 16, 256)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_128_1_conv2d_conv2d  (None, 16, 16, 512)       1180160
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_128_1_conv2d_leakyre (None, 16, 16, 512)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_128_1_pixelshuffler  (None, 32, 32, 128)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_64_1_conv2d_conv2d ( (None, 32, 32, 256)       295168
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_64_1_conv2d_leakyrel (None, 32, 32, 256)       0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  upscale_64_1_pixelshuffler ( (None, 64, 64, 64)        0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 VERBOSE  face_out_b_conv2d (Conv2D)   (None, 64, 64, 3)         4803
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________Reading training images (A):  61%|██████    | 1186/1959 [00:01<00:05, 138.64it/s]

09/07/2020 22:20:31 VERBOSE  face_out_b (Activation)      (None, 64, 64, 3)         0
09/07/2020 22:20:31 VERBOSE  =================================================================
09/07/2020 22:20:31 VERBOSE  Total params: 6,199,747
09/07/2020 22:20:31 VERBOSE  Trainable params: 6,199,747
09/07/2020 22:20:31 VERBOSE  Non-trainable params: 0
09/07/2020 22:20:31 VERBOSE  _________________________________________________________________
09/07/2020 22:20:31 INFO     Loading Trainer from Original plugin...
09/07/2020 22:20:31 VERBOSE  Loading config: '/Users/name/faceswap/config/train.ini'
Reading training images (A):  66%|██████▌   | 1296/1959 [00:01<00:03, 179.82it/s]
Reading training images (A):  71%|███████   | 1384/1959 [00:01<00:02, 227.71it/s]
Reading training images (A):  81%|████████  | 1586/1959 [00:01<00:01, 298.45it/s]
Reading training images (A):  86%|████████▌ | 1675/1959 [00:01<00:00, 345.28it/s]


Reading training images (B):   0%|          | 0/2621 [00:00<?, ?it/s]
Reading training images (B):   0%|          | 9/2621 [00:00<00:33, 77.78it/s]
Reading training images (B):   1%|          | 21/2621 [00:00<00:30, 86.30it/s]
Reading training images (B):  18%|█▊        | 467/2621 [00:00<00:17, 121.24it/s]
Reading training images (B):  21%|██        | 556/2621 [00:00<00:12, 163.04it/s]
Reading training images (B):  25%|██▍       | 644/2621 [00:00<00:09, 215.36it/s]
Reading training images (B):  29%|██▊       | 753/2621 [00:00<00:06, 282.64it/s]
Reading training images (B):  39%|███▉      | 1028/2621 [00:00<00:04, 386.69it/s]
Reading training images (B):  45%|████▍     | 1171/2621 [00:01<00:03, 400.30it/s]
Reading training images (B):  49%|████▉     | 1284/2621 [00:01<00:02, 446.37it/s]
Reading training images (B):  75%|███████▍  | 1959/2621 [00:01<00:01, 619.43it/s]
Reading training images (B):  85%|████████▍ | 2222/2621 [00:01<00:00, 662.93it/s]
09/07/2020 22:20:35 VERBOSE  Alignments filepath: '/Volumes/nameUSB/DF/faceswap-master/Kevin/Kevin_alignments.fsa'
09/07/2020 22:20:35 INFO     Reading alignments from: '/Volumes/nameUSB/DF/faceswap-master/Kevin/Kevin_alignments.fsa'
09/07/2020 22:20:45 VERBOSE  Alignments filepath: '/Volumes/nameUSB/DF/faceswap-master/TheRock/alignments_merged_20200907_041113.fsa'
09/07/2020 22:20:45 INFO     Reading alignments from: '/Volumes/nameUSB/DF/faceswap-master/TheRock/alignments_merged_20200907_041113.fsa'

2020-09-07 22:22:07.922468: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA09/07/2020 22:22:07 VERBOSE  Enabled TensorBoard Logging

2020-09-07 22:22:07.935791: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7fc5165e20a0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-09-07 22:22:07.935806: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
09/07/2020 22:22:11 INFO     Analyzing Ops: 461 of 1468 operations complete
09/07/2020 22:22:13 INFO     Analyzing Ops: 893 of 1468 operations complete
Process exited.

Re: Faceswap Gui Aborts?

Posted: Mon Sep 07, 2020 10:21 pm
by bryanlyon

Intel - Iris Pro -- This GPU is not supported. If you have an AMD card, you might need to use "plaidml-setup" to select the AMD card. If you have an Nvidia card you need to reinstall and make sure you select the Nvidia version. If you don't have either one, you'll have to use CPU mode which wont train very fast.