v0.2.0b1

[sonichi]

This is a beta release of v0.2.0.

Highlights

• Switching to openai v1. Please read the migration guide and report bugs.
• Support async function execution & get_human_input.
• Improvements in documentation and notebooks.

Thanks to all the reviewers for the v0.2 migration. Thanks to @aayushchhabra1999 @bonadio @marcgreen and other contributors!
 

What's Changed

• Update Installation.md by @gagb in Update Installation.md #456
• Update FAQ with workaround for Issue chromadb fails in codespaces because of old version of sqlite3 #251 by @marcgreen in Update FAQ with workaround for Issue #251 #405
• Fix typo in README.md by @eltociear in Fix typo in README.md #481
• Fix/async function and tool execution by @aayushchhabra1999 in Fix/async function and tool execution #87
• Adding async support to get_human_input by @bonadio in Adding async support to get_human_input #466
• Added example .txt file for agentchat_langchain sample notebook by @jasondotparse in Added example .txt file for agentchat_langchain sample notebook #373
• Fix : Typo by @AaadityaG in Fix : Typo  #506
• Dev/v0.2 by @sonichi in Dev/v0.2 #393

New Contributors

• @marcgreen made their first contribution in Update FAQ with workaround for Issue #251 #405
• @aayushchhabra1999 made their first contribution in Fix/async function and tool execution #87
• @bonadio made their first contribution in Adding async support to get_human_input #466
• @jasondotparse made their first contribution in Added example .txt file for agentchat_langchain sample notebook #373
• @AaadityaG made their first contribution in Fix : Typo  #506

Full Changelog: v0.1.14...v0.2.0b1

---

This discussion was created from the release v0.2.0b1.

[kfsone]

Will this in any way make it easier to target custom OpenAI v1 API endpoints, such as presented by ollama and LM studio?

If you haven't seen LM studio, (windows approach) drop the installer into a Windows Sandbox (optional feature but available in all WIndows 10/11), install, run, point at a model, try the chat, tab on the left, try the server tab on the left, paste the example CURL command line.

ollama is basically the os-image concept of docker recognizing the LM models are .. images.

I'm personally more interested in trying GCP's Vertex models, especially bison. Using test chats via the GCP console with it produces some really exciting results for coding tasks, it blows bard out of the water in comparison. It initiates some wrong answers if you challenge it with predictable edge cases (domain-specific niches with unanswered questions on S/O or Quora) resulting in course corrections I've never seen in Bard or GP4: (paraphrasing) "Square the fox, except fox isn't a number, so lets assume I mean't the irrational number i". Feels like they perhaps trained it on transcripts of programming talks.

[sonichi]

There should be no change in the difficulty of doing that per my understanding. I could be wrong.