MrDeepFakes Forums
  • New and improved dark forum theme!
  • Guests can now comment on videos on the tube.
   
  •  Previous
  • 1
  • 2
  • 3
  • 4(current)
  • 5
  • 6
  • ...
  • 26
  • Next 
DeepFaceLab - Google Colab Tutorial
#31
I've been training with 2 accounts with the iteration now at 68K. just faced this error:

Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).
Time to backup: 11 hours
/content
Running trainer.

Error: expected str, bytes or os.PathLike object, not NoneType
Traceback (most recent call last):
File "/content/DeepFaceLab/mainscripts/Trainer.py", line 22, in trainerThread
pretraining_data_path = Path( args.get('pretraining_data_dir', '') )
File "/usr/lib/python3.6/pathlib.py", line 1001, in __new__
self = cls._from_parts(args, init=False)
File "/usr/lib/python3.6/pathlib.py", line 656, in _from_parts
drv, root, parts = self._parse_args(args)
File "/usr/lib/python3.6/pathlib.py", line 640, in _parse_args
a = os.fspath(a)
TypeError: expected str, bytes or os.PathLike object, not NoneType
Done.
===BOUNTY===
I need a .bat file which takes the aligned folder and masks the whole 256x256 aligned dst image. I'll make a deepfake video of your choice for FREE if you can create/get me that file.
!nvidia-smi
#32
@tania01

DFL has been updated and so I need to make changes to the code on Colab. I will update DFL-Colab in a few hours. You do not need to do anything, just watch when the update comes out.

You can still watch the changelog DFL, because there is a new option in SAE.
#33
Updated DFL-Colab. Now the "Training" is working again. In addition, you can now display previews directly to the output. Thanks for the code LuckyGeorge75

If you have any errors, write here or in PM
#34
[SOLVED]

Hi, I made a copy of my workspace folder with just /model /data_src/aligned and /data_dst/aligned because I just want to do the training.

It has been zipped and put in my root folder,

When I start to train my model I get this :


Code:
Loading: 100% 823/823 [00:01<00:00, 606.12it/s]
Loading: 100% 19377/19377 [00:25<00:00, 746.51it/s]
===== Model summary =====
== Model name: SAE
==
== Current iteration: 0
==
== 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 : False
== |== ca_weights : False
== |== pixel_loss : False
== |== face_style_power : 0.0
== |== bg_style_power : 0.0
== Running on:
== |== [0 : Tesla T4]
=========================
Starting. Press "Enter" to stop training and save model.
[23:19:53][#000001][12.13s][1.6570][1.0832]
Done.
/usr/lib/python3.6/multiprocessing/semaphore_tracker.py:143: UserWarning: semaphore_tracker: There appear to be 1 leaked semaphores to clean up at shutdown
 len(cache))
/usr/lib/python3.6/multiprocessing/semaphore_tracker.py:143: UserWarning: semaphore_tracker: There appear to be 1 leaked semaphores to clean up at shutdown
 len(cache))
/usr/lib/python3.6/multiprocessing/semaphore_tracker.py:143: UserWarning: semaphore_tracker: There appear to be 1 leaked semaphores to clean up at shutdown
 len(cache))
/usr/lib/python3.6/multiprocessing/semaphore_tracker.py:143: UserWarning: semaphore_tracker: There appear to be 1 leaked semaphores to clean up at shutdown
 len(cache))
#35
@chervonij does the new version still have 11 hour auto "workspace" backup enabled? and i don't see any auto "model" save either (where it saves the model every ~15mins and goes down a line)
===BOUNTY===
I need a .bat file which takes the aligned folder and masks the whole 256x256 aligned dst image. I'll make a deepfake video of your choice for FREE if you can create/get me that file.
!nvidia-smi
#36
(05-02-2019, 06:02 AM)tania01 Wrote: You are not allowed to view links. Register or Login to view.@chervonij does the new version still have 11 hour auto "workspace" backup enabled? and i don't see any auto "model" save either (where it saves the model every ~15mins and goes down a line)

Currently no need for newer version since DFL reverted back. Yes it will always have 11 hour auto workspace backup.

It does not have model backup. Just don't use pixel loss.
~ Fake it till you make it ~
You are not allowed to view links. Register or Login to view.
#37
Thanks for this, works so much better than on my own pc.
Do not bother trying to create fakes on an rx580, what took me a day is done in an hour using collab
#38
I have this error, not sure how to resolve:
Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).
Time to backup: 11 hours
/content
Running trainer.

Training data src directory does not exist.
Done.
#39
(05-04-2019, 01:40 AM)biggityballsack Wrote: You are not allowed to view links. Register or Login to view.I have this error, not sure how to resolve:
Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).
Time to backup: 11 hours
/content
Running trainer.

Training data src directory does not exist.
Done.

You need to import your workspace with everything ready to train.
~ Fake it till you make it ~
You are not allowed to view links. Register or Login to view.
#40
@chervonij How do I properly use the export to URL cell?

I extracted images in PNG and I don't have enough space on drive to download it.
~ Fake it till you make it ~
You are not allowed to view links. Register or Login to view.
  •  Previous
  • 1
  • 2
  • 3
  • 4(current)
  • 5
  • 6
  • ...
  • 26
  • Next 

Forum Jump:

Users browsing this thread: 2 Guest(s)