MrDeepFakes Forums

Some content may not be available to Guests. Consider registering an account to enjoy unrestricted access to guides, support and tools

  • We are looking for community members who are intested in helping out. See our HELP WANTED post.

FAN masks?

How to get the FAN mask options to work on openCL version?  Using download version from 5/20/2019.  Tried all and get errors every time.


Running converter.

Loading model...
Using plaidml.keras.backend backend.
INFO:plaidml:Opening device "opencl_amd_tonga.0"
===== Model summary =====
== Model name: SAE
==
== Current iteration: 360
==
== Model options:
== |== batch_size : 4
== |== sort_by_yaw : False
== |== random_flip : True
== |== resolution : 128
== |== face_type : f
== |== learn_mask : True
== |== optimizer_mode : 1
== |== archi : df
== |== ae_dims : 512
== |== e_ch_dims : 42
== |== d_ch_dims : 21
== |== multiscale_decoder : True
== |== ca_weights : True
== |== pixel_loss : False
== |== face_style_power : 0.0
== |== bg_style_power : 0.0
== |== apply_random_ct : False
== Running on:
== |== [0 : Advanced Micro Devices, Inc. Tonga (OpenCL)]
=========================
INFO:plaidml:Analyzing Ops: 84 of 362 operations complete
INFO:plaidml:Analyzing Ops: 119 of 362 operations complete
INFO:plaidml:Analyzing Ops: 153 of 362 operations complete
INFO:plaidml:Analyzing Ops: 184 of 362 operations complete
INFO:plaidml:Analyzing Ops: 218 of 362 operations complete
INFO:plaidml:Analyzing Ops: 276 of 362 operations complete
INFO:plaidml:Analyzing Ops: 328 of 362 operations complete
Choose mode: (1) overlay, (2) hist match, (3) hist match bw, (4) seamless, (5) r
aw. Default - 4 :
4
Seamless hist match? (y/n skip:n) :
n
Mask mode: (1) learned, (2) dst, (3) FAN-prd, (4) FAN-dst , (5) FAN-prd*FAN-dst
(6) learned*FAN-prd*FAN-dst (?) help. Default - 1 : 4
Choose erode mask modifier [-200..200] (skip:0) :
0
Choose blur mask modifier [-200..200] (skip:0) :
0
Choose output face scale modifier [-50..50] (skip:0) :
0
Apply color transfer to predicted face? Choose mode ( rct/lct skip:None ) :
None
Apply super resolution? (y/n ?:help skip:n) :
n
Degrade color power of final image [0..100] (skip:0) :
0
Export png with alpha channel? (y/n skip:n) :
n

Collecting alignments: 0it [00:00, ?it/s]
Running on CPU0.
Using plaidml.keras.backend backend.
Exception: Traceback (most recent call last):
  File "C:\DeepFaceLabOpenCLSSE\_internal\DeepFaceLab\joblib\SubprocessorBase.py
", line 59, in _subprocess_run
    self.on_initialize(client_dict)
  File "C:\DeepFaceLabOpenCLSSE\_internal\DeepFaceLab\mainscripts\Converter.py",
 line 47, in on_initialize
    self.converter.on_cli_initialize()
  File "C:\DeepFaceLabOpenCLSSE\_internal\DeepFaceLab\converters\ConverterMasked
.py", line 128, in on_cli_initialize
    self.fan_seg = FANSegmentator(256, FaceType.toString( self.face_type ) )
  File "C:\DeepFaceLabOpenCLSSE\_internal\DeepFaceLab\facelib\FANSegmentator.py"
, line 25, in __init__
    exec( nnlib.import_all(), locals(), globals() )
  File "C:\DeepFaceLabOpenCLSSE\_internal\DeepFaceLab\nnlib\nnlib.py", line 754,
 in import_all
    nnlib.import_keras(device_config)
  File "C:\DeepFaceLabOpenCLSSE\_internal\DeepFaceLab\nnlib\nnlib.py", line 188,
 in import_keras
    nnlib.keras.backend.set_session(nnlib.tf_sess)
AttributeError: module 'keras.backend' has no attribute 'set_session'
 
Can someone help me understand if I'm doing something wrong, or what the issue is?  I can use learned mask, but all of the FAN options give errors.  How exactly are the FAN mask options used and why can't I use them?
 

helpplease

DF Vagrant
TheMadFaker said:
Can someone help me understand if I'm doing something wrong, or what the issue is?  I can use learned mask, but all of the FAN options give errors.  How exactly are the FAN mask options used and why can't I use them?

lol i think im having the exact same problem as you, and i just posted about it before i saw your post. i hope someone can help us :s
 
I guess it's not just and AMD issue then as you are using the CUDA version right? I'm still not able to use the FAN options at conversion. I don't understand if there's something else I have to do before it will work properly. Is there something more to it than just choosing one of the options when converting?
 
Top