More complete initialization of agents Thanks to @jxnlco you can now more completely initialize agents by specifying `agent_kwargs` This allows for passing of more complex parameters to the agent class https://
github.com/hwchase17/lang
chain/blob/b9045f7e0df357d608700e91dc3039433f652d61/langchain/agents/initialize.py#L17
…
@langchain
-
Enhanced Agent Initialization with Advanced Parameters
By
–
-
LangChain v0.0.75 Release: Agents, Embeddings, SQL Steps
By
–
Another big release! v0.0.75 Pin LangChainHub version (
@roywilliamsiii
)
TensorflowHub embeddings (Ryohei Kuroki)
Returning intermediate SQL steps
Better initialization of agents (
@jxnlco
)
Using MMR search in the vector DB QA chain And more! -
LangChainHub Now Supports Git Commit Version Pinning
By
–
Pin LangChainHub version Thanks to @roywilliamsiii you can now load artifacts from LangChainHub from a specific git commit, effectively pinning the version Docs: https://
langchain.readthedocs.io/en/latest/modu
les/agents/examples/load_from_hub.html#pinning-dependencies
… Check out the Hub here: -
Best Practices for Deploying LangChain Applications
By
–
An awesome example for everyone asking how to best deploy langchain apps! https://t.co/GSrE53HYgj
— LangChain (@LangChain) 30 janvier 2023An awesome example for everyone asking how to best deploy langchain apps!
-
LangChain Introduces Tracing Feature for Chain Debugging
By
–
We’re excited to announce a new feature in LangChain: Tracing! By enabling tracing in LangChain, you’ll be able to more effectively inspect and debug your complex chain/agent runs. Blog: https://
blog.langchain.dev/tracing/
Docs: https://
langchain.readthedocs.io/en/latest/trac
ing.html
… -
LangChain Apps Moving to Production Support Channel
By
–
Hearing from a few folks that apps built on LangChain are moving into production! If that's the case for you, we'd love to set up a Slack connect channel to offer better support. Fill out the form below and we'll set it up! https://
forms.gle/57d8AmXBYp8PP8
tZA
… If you don't use Slack, DM us -
LangChain Updates: Context Windows, Cohere Support, Bug Fixes
By
–
Other improvements: – Correct context window for davinci-003 :
@wskish – Docs improvements: wunderalbert
– Fix bash util: h/t Marcin in discord
– Add support for arbitrary params to Cohere wrapper: h/t bair82 in discord
– Fix caching bug: @brucehammer -
SQLDatabaseChain improvements with better data understanding
By
–
Improvements to SQL database chain @fpingham with some key improvements to SQLDatabaseChain 1. Optionally include first row in prompt (for better understanding of data) Docs: https://
langchain.readthedocs.io/en/latest/modu
les/chains/examples/sqlite.html#adding-first-row-of-each-table
… 2. Improve default prompt -
LangChain Search Classes Enhanced with Detailed Metadata Results
By
–
More detailed search results @EnricoShippole added a method to the search classes to return more detailed search info: title, snippet, link WebGPT? Google Search: https://
langchain.readthedocs.io/en/latest/modu
les/utils/examples/google_search.html#metadata-results
… Bing Search: https://
langchain.readthedocs.io/en/latest/modu
les/utils/examples/bing_search.html#metadata-results
… -
Code improvements and agent refactoring for cleaner codebase
By
–
A few code improvements from Amos Ng – Improvement of some tests for debugging purposes
– Refactor of agents to cleanup the base class Both go a long way in keeping a clean codebase