0.0.20

What's Changed

• update arm docker by @michaelfeil in #73
• patch release: optimum tokenization issue

Full Changelog: 0.0.19...0.0.20

0.0.19 - yanked

What's Changed

• update optimum_utils by @michaelfeil in #72

Full Changelog: 0.0.18...0.0.19

0.0.18 - yanked

What's Changed

• support mps backend. by @ninehills in #59
• Add optimum[onnx] by @michaelfeil in #68

New Contributors

• @ninehills made their first contribution in #59 Thanks @ninehills for sharing this on twitter.

Full Changelog: 0.0.17...0.0.18

0.0.17

What's Changed

Breaking: Switched to Cuda 12.1 and torch 2.1.2

• Add rerank/predict endpoint in the API by @michaelfeil in #50
• update dockerfile (Cuda 12.1 and torch 2.1.2) and tests by @michaelfeil in #54

Full Changelog: 0.0.16...0.0.17

0.0.16

What's Changed

• fixing delayed warmup by @michaelfeil in #53
• expose capabilities by @michaelfeil in #53

Full Changelog: 0.0.15...0.0.16

0.0.15

What's Changed

• Linting and better model errors by @michaelfeil in #52

Full Changelog: 0.0.14...0.0.15

0.0.14

What's Changed

• adding tests and better typing by @michaelfeil in #51

Full Changelog: 0.0.13...0.0.14

0.0.13

What's Changed

• update-ci-and-readme by @michaelfeil in #48
• fix-fp16 by @michaelfeil in #49

Full Changelog: 0.0.12...0.0.13

0.0.12

What's Changed

• add classification pipeline by @michaelfeil in #44 -> See update in Readme.md
• make fp16 default for CUDA by @michaelfeil in #47
• add bettertransformer by @michaelfeil in #47

Breaking:

• make fp16 default for CUDA by @michaelfeil in #47 -> This may lower your precision for a 2x speedup.

Full Changelog: 0.0.11...0.0.12

0.0.11

What's Changed

• Refactoring of API. by @michaelfeil in #43
• initial support for ReRanker models by @michaelfeil in #43

Full Changelog: 0.0.10...0.0.11