Page 1 of 1

Problem with Model directory

Posted: Sat May 30, 2020 5:43 pm
by dinarama8

Hello,

I trained the system without any problem, but when I go to the "convert" step, I am not able as I always got this message: "Model could not be found in folder 'C:\Users\mturr\Videos\Faceswap\Prueba\Model'" . This is the model directory used in the training step, and contains:

  • original_decoder_A.h5
  • original_decoder_B.h5
  • original_encoder.h5
  • original_state.json
    and original_logs directory and several .bk files

What does it miss? What am I doing wrong?

Thanks and best,

Miguel


Re: Problem with Model directory

Posted: Sun May 31, 2020 9:18 am
by torzdf

Can you please post the output from the "Generate" button with your convert settings, and a screengrab of your model folder (including the address bar)


Convert Function - Model could not be found ERROR

Posted: Sat Jun 06, 2020 9:22 am
by RahulRookie

Dear Experts, need some guidance

I am creating my video using latest version of Faceswap application

After Training the model, when I run Convert function, I am getting the following error message

Code: Select all

06/06/2020 14:31:01 ERROR    Model could not be found in folder 'D:\FACESWAP_PROJECTS\PROJECT_1\ModelAB'. Exiting

Need your help to know if I am doing anything wrong and how to fix this error, why it gives this error message when all the files present?

Thanks for your help

Output from Generate command

C:\Users\Rahul\MiniConda3\envs\faceswap\python.exe D:\faceswap\faceswap.py convert -i D:/FACESWAP_PROJECTS/PROJECT_1/GENELIA.mov -o D:/FACESWAP_PROJECTS/PROJECT_1/Output -m D:/FACESWAP_PROJECTS/PROJECT_1/ModelAB -c none -M none -sc none -w ffmpeg -osc 100 -l 0.4 -j 0 -t dfaker -otf -L INFO

Directory of D:\FACESWAP_PROJECTS\PROJECT_1\ModelAB


06-06-2020 02.44 PM <DIR> .
06-06-2020 02.44 PM <DIR> ..
04-06-2020 08.56 AM 24,828,816 original_decoder_A.h5
04-06-2020 08.14 AM 24,828,816 original_decoder_A.h5.bk
04-06-2020 08.56 AM 24,828,816 original_decoder_B.h5
04-06-2020 08.14 AM 24,828,816 original_decoder_B.h5.bk
04-06-2020 08.56 AM 278,691,768 original_encoder.h5
04-06-2020 08.14 AM 278,691,768 original_encoder.h5.bk
04-06-2020 08.58 AM <DIR> original_logs
04-06-2020 08.56 AM 922 original_state.json
04-06-2020 08.14 AM 921 original_state.json.bk


Re: Convert Function - Model could not be found ERROR

Posted: Sat Jun 06, 2020 9:48 am
by torzdf

I have never seen this before, and this looks fine on the face of it....

Triple check your paths, and check permissions on the model folder?


Re: Convert Function - Model could not be found ERROR

Posted: Sat Jun 06, 2020 8:20 pm
by Tekniklee

Just a note on this - the field can contain the correct file path and name, but you need to use the browse and select to get the file. Many times (such as switching from initial to final training sets) I have tried to just change the path or filename in the box. However, when the model runs, you get not found errors until you browse and select the actual file or folder. It's not like it tries to run using the old contents, but rather as though any manual update to the field sets it as no contents. As far as I know, this behavior is true on all screens, even things like log and timelapse files. Maybe capitalization sensitivity? Anyway, browse and select your model and other items it's not finding.

As a side note, it would be really nice to change the mouse hover on fields from item help to field contents, so you can just hover over a field to see it's full path, versus having to click on it and run the cursor to the left and right.