CalvinBackup/neuralchen-SimSwap
1
0
Fork 0
Code Issues 288 Pull Requests 16 Actions Packages Projects Releases 2 Wiki Activity

Python and Pytorch Version for Train #250

New Issue
Open
opened 2022-04-28 07:43:41 +02:00 by papipulato · 3 comments
papipulato commented 2022-04-28 07:43:41 +02:00 (Migrated from github.com)
Copy Link
Copy Source

i have newbie question, before update on train model im using insightface 0.2.1 python 3.9 and january 2022 pytorch. its still work for swap video (no error) and image (error word but still have good result). then pytorch update on lastest, got all of error on them. iam using aws cloud gpu maybe its default to python 3.9.

  • how to back my python and pytorch to 3.6 ?
  • can i use python 3.9 and january 2022 pytorch to train new method model ? i will comeback with log error.
    thanks
i have newbie question, before update on train model im using insightface 0.2.1 python 3.9 and january 2022 pytorch. its still work for swap video (no error) and image (error word but still have good result). then pytorch update on lastest, got all of error on them. iam using aws cloud gpu maybe its default to python 3.9. - how to back my python and pytorch to 3.6 ? - can i use python 3.9 and january 2022 pytorch to train new method model ? i will comeback with log error. thanks
ExponentialML commented 2022-04-28 08:22:11 +02:00 (Migrated from github.com)
Copy Link
Copy Source

i have newbie question, before update on train model im using insightface 0.2.1 python 3.9 and january 2022 pytorch. its still work for swap video (no error) and image (error word but still have good result). then pytorch update on lastest, got all of error on them. iam using aws cloud gpu maybe its default to python 3.9.

  • how to back my python and pytorch to 3.6 ?
  • can i use python 3.9 and january 2022 pytorch to train new method model ? i will comeback with log error.
    thanks

If unsure about how to install, always run using the preparation documentation

Using Anaconda allows you to install whatever version of Python you want in a "container" environment. That way, you don't have to mess around with AWS version of Python.
Also, if you have errors with Pytorch, it's most likely that version is not compatible with the GPU you're using. You have to search for a compatible version here.
For example, the newer GPUs like the 3000 series don't support CUDA 10.1, so you have to install 11.1 when installing torch. Make sure your Pytorch version matches the one in the preparation documentation.

> i have newbie question, before update on train model im using insightface 0.2.1 python 3.9 and january 2022 pytorch. its still work for swap video (no error) and image (error word but still have good result). then pytorch update on lastest, got all of error on them. iam using aws cloud gpu maybe its default to python 3.9. > > * how to back my python and pytorch to 3.6 ? > * can i use python 3.9 and january 2022 pytorch to train new method model ? i will comeback with log error. > thanks If unsure about how to install, always run using the [preparation documentation](https://github.com/neuralchen/SimSwap/blob/main/docs/guidance/preparation.md) Using Anaconda allows you to install whatever version of Python you want in a "container" environment. That way, you don't have to mess around with AWS version of Python. Also, if you have errors with Pytorch, it's most likely that version is not compatible with the GPU you're using. You have to search for a compatible version [here](https://pytorch.org/get-started/previous-versions/). For example, the newer GPUs like the 3000 series don't support CUDA 10.1, so you have to install 11.1 when installing torch. Make sure your Pytorch version matches the one in the preparation documentation.
👍 2
neuralchen commented 2022-04-28 08:24:55 +02:00 (Migrated from github.com)
Copy Link
Copy Source

The main part of new available training scripts have no relationship with insightface respo.
Only the face detection function depend on insightface.
Face detection is only performed when processing the unaligned picture.

The main part of new available training scripts have no relationship with insightface respo. Only the face detection function depend on insightface. Face detection is only performed when processing the unaligned picture.
👍 1
Fibonacci134 commented 2022-04-29 16:47:40 +02:00 (Migrated from github.com)
Copy Link
Copy Source

i have newbie question, before update on train model im using insightface 0.2.1 python 3.9 and january 2022 pytorch. its still work for swap video (no error) and image (error word but still have good result). then pytorch update on lastest, got all of error on them. iam using aws cloud gpu maybe its default to python 3.9.

  • how to back my python and pytorch to 3.6 ?
  • can i use python 3.9 and january 2022 pytorch to train new method model ? i will comeback with log error.
    thanks

A few tips:
-Try not to mix and match your cuda versions.
-Make sure you're installing pytorch with cuda, from the pytorch official website.

  • Always have a backup of your environment!
    -You might even want to set up a different environment for training Incase you're setting up tensorRt.

Also, do this in terminal:
python
Import torch
torch.cuda.is_available()

-This should return true. You might be having compatibility issues with python 3.9; in which case just create a new environment with python 3.6 or 3.7

Good luck.

> i have newbie question, before update on train model im using insightface 0.2.1 python 3.9 and january 2022 pytorch. its still work for swap video (no error) and image (error word but still have good result). then pytorch update on lastest, got all of error on them. iam using aws cloud gpu maybe its default to python 3.9. > - how to back my python and pytorch to 3.6 ? > - can i use python 3.9 and january 2022 pytorch to train new method model ? i will comeback with log error. > thanks A few tips: -Try not to mix and match your cuda versions. -Make sure you're installing pytorch with cuda, from the pytorch official website. - Always have a backup of your environment! -You might even want to set up a different environment for training Incase you're setting up tensorRt. Also, do this in terminal: python Import torch torch.cuda.is_available() -This should return true. You might be having compatibility issues with python 3.9; in which case just create a new environment with python 3.6 or 3.7 Good luck.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
GIF
GUI
Increase inference time
A solution to increase inference time
 bug
Something isn't working
 documentation
Improvements or additions to documentation
 duplicate
This issue or pull request already exists
 enhancement
New feature or request
 good first issue
Good for newcomers
 help wanted
Extra attention is needed
 invalid
This doesn't seem right
 question
Further information is requested
 wontfix
This will not be worked on
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
CalvinBackup (Calvin Backup)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: CalvinBackup/neuralchen-SimSwap#250
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?