Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Don't reload models everytime convert_one_image is called #39

Closed
Ganonmaster opened this issue Dec 28, 2017 · 3 comments
Closed

Don't reload models everytime convert_one_image is called #39

Ganonmaster opened this issue Dec 28, 2017 · 3 comments
Assignees
@Clorr
Labels
bug dev good first issue performance
Projects
Usability
Milestone
First pre-release

Comments

@Ganonmaster
Copy link
Contributor

Expected behavior

Use the convert command to convert a directory. convert_one_image loads the model once.

Actual behavior

Use the convert command to convert a directory. convert_one_image loads the model every time that it is called.
@Ganonmaster Ganonmaster mentioned this issue Dec 28, 2017
Closed
@Ganonmaster Ganonmaster added this to To Do in Usability Dec 30, 2017
@Ganonmaster Ganonmaster added this to the First pre-release milestone Dec 30, 2017
@Clorr Clorr self-assigned this Jan 1, 2018
@Clorr
Copy link
Contributor

Clorr commented Jan 1, 2018 •
edited

I'm adding a Convert class which will load models on __init__

It will be in my branch until I do a PR, but feel free to have a look

@Clorr
Copy link
Contributor

Clorr commented Jan 1, 2018 •
edited

The class is in my PR as the Convert_Adjust.py plugin. However, I did not address the fact that the model is reloaded for each image.

@Ganonmaster , let me know if you want to tackle this separately, or if you are ok with my plugins folder. If so I'll make it run so it can be integrated quickly

@deepfakes deepfakes mentioned this issue Jan 2, 2018
Closed
Clorr added a commit to Clorr/faceswap that referenced this issue Jan 5, 2018
@Clorr
Do not reload model on each image deepfakes#39
232e1ec
@Clorr
Copy link
Contributor

Clorr commented Jan 5, 2018

This is handled in PR #52 in a dedicated commit

Clorr added a commit to Clorr/faceswap that referenced this issue Jan 5, 2018
@Clorr
Do not reload model on each image deepfakes#39
7e7c59d
Clorr added a commit to Clorr/faceswap that referenced this issue Jan 5, 2018
@Clorr
Do not reload model on each image deepfakes#39
5c98713
Clorr added a commit to Clorr/faceswap that referenced this issue Jan 5, 2018
@Clorr
Do not reload model on each image deepfakes#39
9dd4cf0
@Clorr Clorr closed this as completed Jan 5, 2018
Usability automation moved this from To Do to Done Jan 5, 2018
Clorr added a commit to Clorr/faceswap that referenced this issue Jan 5, 2018
@Clorr
Do not reload model on each image deepfakes#39
7139584
Clorr added a commit to Clorr/faceswap that referenced this issue Jan 8, 2018
@Clorr
Do not reload model on each image deepfakes#39
105bd33
Clorr added a commit to Clorr/faceswap that referenced this issue Jan 8, 2018
@Clorr
Do not reload model on each image deepfakes#39 + Adding FaceFilter de…
544954f 
…epfakes#53
Clorr added a commit that referenced this issue Jan 31, 2018
@Clorr
Adding models as plugins + Face filtering (#53) + #39 + #43 + #44 + #49…
34945cf 
… (#61)

* Making Models as plugins

* Do not reload model on each image #39 + Adding FaceFilter #53

* Adding @lukaville PR for #43 and #44 (possibly)

* Training done in a separate thread

* Better log for plugin load

* Adding a prefetch to train.py #49
(Note that we prefetch 2 batches of images, due to the queue behavior)
+ More compact logging with verbose info included

* correction of DirectoryProcessor signature

* adding missing import

* Convert with parallel preprocessing of files

* Added coverage var for trainer

Added a var with comment. Feel free to add it as argument

* corrections

* Modifying preview and normalization of image + correction

* Cleanup
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@Clorr Clorr

Labels
bug dev good first issue performance
Projects
No open projects
Usability
  
Done
Milestone
First pre-release
Development

No branches or pull requests
2 participants
@Ganonmaster @Clorr