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.

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

Version 0.8.6 #1212

Merged
merged 98 commits into from
Apr 7, 2024
Merged

Version 0.8.6 #1212

merged 98 commits into from
Apr 7, 2024

Conversation

kohya-ss
Copy link
Owner

  • Update PyTorch and dependent libraries
  • DeepSpeed
  • Masked loss
  • Dataset info cache
  • Wildcard, multiline, second separator etc. for captions.
  • wd14 tagger v3 support
  • etc.

kohya-ss and others added 29 commits March 25, 2024 22:28
Add metadata caching for DreamBooth dataset
Add OpenVINO and ROCm ONNX Runtime for WD14
Rating support for WD Tagger
verify command line args if wandb is enabled
…mprove resilience to data corruption (#1228)

* add huber loss and huber_c compute to train_util

* add reduction modes

* add huber_c retrieval from timestep getter

* move get timesteps and huber to own function

* add conditional loss to all training scripts

* add cond loss to train network

* add (scheduled) huber_loss to args

* fixup twice timesteps getting

* PHL-schedule should depend on noise scheduler's num timesteps

* *2 multiplier to huber loss cause of 1/2 a^2 conv.

The Taylor expansion of sqrt near zero gives 1/2 a^2, which differs from a^2 of the standard MSE loss. This change scales them better against one another

* add option for smooth l1 (huber / delta)

* unify huber scheduling

* add snr huber scheduler

---------

Co-authored-by: Kohya S <[email protected]>
@kohya-ss kohya-ss merged commit 683f3d6 into main Apr 7, 2024
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

10 participants