chatify

First release on pip!

What's Changed

• prompt tweaks by @jeremymanning in https://github.com/ContextLab/chatify/pull/41
• Added RequestChain by @HemuManju in https://github.com/ContextLab/chatify/pull/43
• lots of prompt tweaks by @jeremymanning in https://github.com/ContextLab/chatify/pull/44
• more prompt tweaks by @jeremymanning in https://github.com/ContextLab/chatify/pull/45
• updated prompts by @jeremymanning in https://github.com/ContextLab/chatify/pull/46
• improved prompts by @jeremymanning in https://github.com/ContextLab/chatify/pull/48
• fix formatting for multiple choice question choices by @jeremymanning in https://github.com/ContextLab/chatify/pull/49
• typo fix by @jeremymanning in https://github.com/ContextLab/chatify/pull/50
• codespell: config, workflow (fail if typos introduced) and some typos fixed by @yarikoptic in https://github.com/ContextLab/chatify/pull/51
• Update setup.py by @HemuManju in https://github.com/ContextLab/chatify/pull/52
• getting ready for release on pip by @jeremymanning in https://github.com/ContextLab/chatify/pull/53

New Contributors

• @yarikoptic made their first contribution in https://github.com/ContextLab/chatify/pull/51

Full Changelog: https://github.com/ContextLab/chatify/compare/v0.2.0...v0.2.1

Lots of optimizations and changes!

Summary (auto-generated):

• fix tab order stuff by @jeremymanning in https://github.com/ContextLab/chatify/pull/34
• formatting fixes by @jeremymanning in https://github.com/ContextLab/chatify/pull/35
• Updated LLM model loading mechanism by @HemuManju in https://github.com/ContextLab/chatify/pull/36
• switch to proxy server by default by @jeremymanning in https://github.com/ContextLab/chatify/pull/39
• adds basic syntax highlighting to code snippets by @jeremymanning in https://github.com/ContextLab/chatify/pull/40

Full Changelog: https://github.com/ContextLab/chatify/compare/v0.1.1...v0.2.0

Big update!

Chatify can now run fully locally in iPython/Jupyter notebooks, including in the free tier notebooks on Colaboratory and Kaggle. We've switched from a cached-response model to a fully interactive model (which generates a unique response to each query) based on Meta's Llama 2 model (we use the 7B variant to minimize resource requirements).

We have also tweaked the prior built-in prompts and added lots of useful new prompts.

Initial release (alpha! untested!) of the Chatify package. This release is being generated to create a zenodo link.

Core contributors: @HemuManju, @jeremymanning, @koerding

Full changelog: https://github.com/ContextLab/chatify/commits/v0.1.0