0.0.7

What's Changed

• [REVIEW] Support dict output for models by @VibhuJawa in #98
• Fix loading multiple models by @VibhuJawa in #99
• Fix a minor bug which was causing loading twice by @VibhuJawa in #101

Full Changelog: 0.0.6...0.0.7

0.0.6

CrossFit 0.0.6 Release Notes

We are excited to announce the release of CrossFit 0.0.6! This version brings significant improvements in performance, compatibility, and usability. We've focused on optimizing workflows, enhancing compatibility with the latest tools, and fixing critical bugs to ensure a smoother experience for our users.

What's Changed

• Update github workflow for GPU Tests by @praateekmahajan in #87
• ListColumn constructor for cudf 24.10 + np.nan vs np.NaN by @praateekmahajan in #86
• Deduplicate columns to select in add_keep_cols by @praateekmahajan in #88
• Added ctranslate2 translation example script by @uahmed93 in #83
• Add citation by @VibhuJawa in #89
• Update publish workflow by @raydouglass in #91
• [REVIEW] Fix conda dev environment by @VibhuJawa in #93
• Fix logging of GPU ID by @VibhuJawa in #95
• Fix pytorch memory curve estimation for rmm backed allocator by @VibhuJawa in #94
• Bump Version to 0.6 by @VibhuJawa in #97

New Contributors

• @praateekmahajan made their first contribution in #87
• @uahmed93 made their first contribution in #83
• @raydouglass made their first contribution in #91

Full Changelog: 0.0.5...0.0.6

0.0.5

What's Changed

• Fix build instructions in README.md by @VibhuJawa in #61
• Bump black from 23.10.1 to 24.3.0 in /requirements by @dependabot in #71
• Fix torch cache cleanup issues by @VibhuJawa in #75
• Convert Pandas Series to cuDF Series by @sarahyurick in #76
• Fix memory leak by @VibhuJawa in #80
• Update setup.py by @VibhuJawa in #81

New Contributors

• @dependabot made their first contribution in #71
• @sarahyurick made their first contribution in #76

Full Changelog: 0.0.4...0.0.5

0.0.4

What's Changed

• Add MANIFEST.in to ensure clean builds by @VibhuJawa in #67
• Fix Padding Related Bugs: Crossfit by @VibhuJawa in #66
• Bump up version to 0.0.4 by @VibhuJawa in #68

Full Changelog: 0.0.3...0.0.4

0.0.3

What's Changed

• Fix CI by @VibhuJawa in #60
• Add pip install commands by @VibhuJawa in #62
• Allow HF and sentence-transformer models by @VibhuJawa in #63
• Update setup.py version by @VibhuJawa in #64

New Contributors

• @VibhuJawa made their first contribution in #60

Full Changelog: 0.0.2...0.0.3

0.0.2

What's Changed

• Port latest code from other repo by @marcromeyn in #1
• Add copy-pr-bot file to enable RAPIDS GPU runners by @oliverholworthy in #3
• Add GPU CI workflow by @jperez999 in #11
• Add GPU CI workflow by @edknv in #8
• add github action for detecting large files by @edknv in #25
• run ci in custom container with model and data by @edknv in #28
• add a unit test for benchmarking by @edknv in #27
• add a github action for benchmarking by @edknv in #26
• skip benchmark comments if merged to main by @edknv in #31
• remove beir dependency by @edknv in #30
• add apache license and attribution by @edknv in #33
• add an example beir report script by @edknv in #35
• fix tokenizer cache file name bug by @edknv in #36
• tear down by unloading from workers by @edknv in #39
• test example scripts in ci by @edknv in #40
• add ops for running a custom pytorch classifier by @edknv in #38
• update setup.py by @edknv in #41
• upgrade base container to nemo:23.10 by @edknv in #42
• add sentence piece tokenizer by @edknv in #43
• update custom pytorch example to use deberta by @edknv in #44
• handle 1d outputs from custom pytorch models by @edknv in #45
• add option to limit max characters by @edknv in #46
• install dberta model in ci docker image by @edknv in #47
• add option to choose axis in labeler by @edknv in #48
• Update dask "compute" benchmark by @rjzamora in #50

New Contributors

• @marcromeyn made their first contribution in #1
• @jperez999 made their first contribution in #11
• @rjzamora made their first contribution in #50

Full Changelog: 0.0.1...0.0.2