Releases: HMUNACHI/nanodl
Releases · HMUNACHI/nanodl
v0.0.0
1.2.5dev1
Full Changelog: 1.2.4dev1...1.2.5dev1
1.2.4dev1
Full Changelog: 1.2.2.dev1...1.2.4dev1
1.2.2.dev1
What's Changed
- Dev by @HMUNACHI in #19
- Syncing by @HMUNACHI in #20
- Tokenizer by @HMUNACHI in #21
- Merge pull request #21 from HMUNACHI/dev by @HMUNACHI in #22
- Dev by @HMUNACHI in #23
- minor bug fix -> Correct Use of dropout probability as Norm epsilon by @danny-1k in #26
- IJEPA implementation by @danny-1k in #28
- touched up Ijepa by @HMUNACHI in #29
- added changes by @HMUNACHI in #30
New Contributors
Full Changelog: 1.2.1.dev1...1.2.2.dev1
What's Changed
- Dev by @HMUNACHI in #19
- Syncing by @HMUNACHI in #20
- Tokenizer by @HMUNACHI in #21
- Merge pull request #21 from HMUNACHI/dev by @HMUNACHI in #22
- Dev by @HMUNACHI in #23
- minor bug fix -> Correct Use of dropout probability as Norm epsilon by @danny-1k in #26
- IJEPA implementation by @danny-1k in #28
- touched up Ijepa by @HMUNACHI in #29
- added changes by @HMUNACHI in #30
New Contributors
Full Changelog: 1.2.1.dev1...1.2.2.dev1
1.2.1.dev1
What's Changed
- Fix instructions to install dependencies in README.md by @afrendeiro in #14
- patches in random by @HMUNACHI in #16
- Fixed random by @HMUNACHI in #17
- Some patches by @HMUNACHI in #18
New Contributors
- @afrendeiro made their first contribution in #14
Full Changelog: 1.2.0.dev1...1.2.1.dev1
NanoDL 1.2.0.dev1
What's Changed
What's New in version 1.2.0.dev1
- Google's Gemma architecture.
- Reward model wrapper and data-parallel distributed reward trainer.
- True random number generators in Jax which do not need the verbose code (examples shown in next sections).
Full Changelog: v1.0.1dev1...1.2.0.dev1
v1.0.1dev1
Dev pre-release
Full Changelog: https://github.com/HMUNACHI/nanodl/commits/v1.0.1dev1