Skip to content

Allow specifying datasets in configs, change default optimizer, improved CUDA support#1

Merged
aronson merged 6 commits into
mainfrom
fizzAI-patch-1
Jan 24, 2026
Merged

Allow specifying datasets in configs, change default optimizer, improved CUDA support#1
aronson merged 6 commits into
mainfrom
fizzAI-patch-1

Conversation

@fizzAI

@fizzAI fizzAI commented Jan 23, 2026

Copy link
Copy Markdown
Member

AdamW is one morbillion times better than Adam

@fizzAI fizzAI changed the title Switch to AdamW, add boilerplate for more optimizers later Switch to AdamW, add boilerplate for more optimizers later, update Python dep to 3.12 Jan 23, 2026
@fizzAI
fizzAI marked this pull request as draft January 23, 2026 23:00
@fizzAI fizzAI changed the title Switch to AdamW, add boilerplate for more optimizers later, update Python dep to 3.12 Allow specifying datasets in configs, change default optimizer, improved CUDA support Jan 23, 2026
@fizzAI
fizzAI marked this pull request as ready for review January 23, 2026 23:39
@aronson
aronson self-requested a review January 23, 2026 23:42

@aronson aronson left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

looks really good, I need to take ~30 minutes to run 2-3 tests on my mac to make sure we still have feature parity. if you could, @fizzAI can you post a screenshot of the application working on your CUDA setup? also name tersely which CUDA version and card you're using for posterity. then I can merge np

Comment thread scripts/mlx_trainer.py Outdated
Comment thread pyproject.toml Outdated
@aronson aronson self-assigned this Jan 23, 2026
@aronson

aronson commented Jan 23, 2026

Copy link
Copy Markdown
Member

LGTM over here

Screen.Recording.2026-01-23.at.5.51.05.PM.mov

@fizzAI

fizzAI commented Jan 23, 2026

Copy link
Copy Markdown
Member Author
image

Nvidia GTX 1650 Super, Cuda 12 MLX and Cuda runtime

@fizzAI
fizzAI requested a review from aronson January 23, 2026 23:59
Signed-off-by: Heni <i@pingas.org>
@aronson
aronson merged commit b1e2523 into main Jan 24, 2026
1 check passed
@aronson
aronson deleted the fizzAI-patch-1 branch January 24, 2026 00:05
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.

2 participants