Utilities to help you work with your language model data outside LangSmith
Project description
langfree
Utilities that help you extract, transform and curate ChatOpenAI runs from traces stored in LangSmith. The goal is to facilitate using data from these runs for fine-tuning / instruction-tuning models.
Motivation
Langchain has native tracing support that allows you to log LangChain runs. This data is a valuable resource for fine-tuning and evaluation. LangSmith is a commercial application that facilitates some of these tasks.
However, LangSmith may not work for everyone’s needs. It is often desirable to buid your own data inspection and curation infrastructure so you can add your own features.
One pattern I noticed is that great AI researchers are willing to manually inspect lots of data. And more than that, they build infrastructure that allows them to manually inspect data quickly. Though not glamorous, manually examining data gives valuable intuitions about the problem. The canonical example here is Andrej Karpathy doing the ImageNet 2000-way classification task himself.
langfree
helps you export data from LangSmith and build data curation
web applications. The goal is to empower the AI engineer to quickly
build their own data curation tools, so they can add features like:
- connectivity to additionaly data sources beyond LangSmith.
- data transformations of runs using code
- ability to route, tag and annotate data in custom ways
- … etc.
Furthermore,langchain
provides a handful of Shiny for
Python components to make the process of
creating data curation applications easier.
Install
pip install langfree
How to use
Coming soon
Documentation
See the docs site.
Contributing
This library was created with nbdev. See Contributing.md for further guidelines.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for langfree-0.0.14-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 015de83406e6185891e29a9acba8ddd8e8718402d943b6938f1495c025ebc033 |
|
MD5 | 0cbb4c8317986fceab446aa88ab725ee |
|
BLAKE2b-256 | 792b5bf570227ca342889c3be1cbf0d1814f2be1132a9a5413ee612a3b089dfb |