Over the weekend, I added a GMail ChatLoader integration to @langchain
, to help anyone to finetune models on their emails I evaluated it the old fashioned way – the eyeball test Here's how it did responding to various types of inquiries (VC, customer, integration, @swyx
)
@hwchase17
-
Gmail ChatLoader Integration for Email Model Finetuning
By
–
-
LangSmith Enables App Development Beyond Python and JavaScript
By
–
If you're building apps not in Python/JS, you can still use LangSmith!
-
Improving SQL Question-Answering Systems with LangChain
By
–
We're continuing to do a deep dive on improving question-answering over SQL data Some initial improvements by @fpingham have already made it much more resilient to questions involving text (song titles, artist names, etc) Access it here: https://
sql-langchain.streamlit.app -
Timeline-Based Data Structure for Personal Event RAG Queries
By
–
Just saw a talk by @AlonHalevy about using a **timeline** based data-structure to do RAG. Really useful for querying over personal events Need to dive into this more! Am always excited to see novel data-structures Repo here: https://
github.com/facebookresear
ch/personal-timeline/tree/main
… -
Fine-tuning LLMs for personalized voice synthesis webinar
By
–
I had a lot of fun hacking on this! Will share more details on this during our webinar on fine-tuning LLMs to talk in your voice this Wednesday!https://t.co/TjYFfy5Ht5 https://t.co/CJKCLM4uWR
— Harrison Chase (@hwchase17) 27 août 2023I had a lot of fun hacking on this! Will share more details on this during our webinar on fine-tuning LLMs to talk in your voice this Wednesday! https://
crowdcast.io/c/lzafugqtyata -
Creative embedding generation techniques for document processing
By
–
The cool thing about this is this can work with arbitrary ways of generating different embeddings for a document E.g. by splitting into smaller chunks, creating a summary of the document, generating hypothetical questions What other creative ways could we do this?
-
New Data Loaders for Fine-tuning AI Models Across Platforms
By
–
We've started adding loaders to help load data in a format such that you can finetune (usually for a custom tone) So far we've got Facebook Messenger, Slack, Telegram, WhatsApp, Twitter (via @apify
), Discord What else should we add? -
LangChain Twitter Fine-tuning Repository Released
By
–
In the github repo! https://
github.com/langchain-ai/t
witter-finetune
… -
Training AI models effectively with minimal labeled data
By
–
was <100 examples, and i actually didnt even have a topic for each tweet…. it seems to work fine without topics (which I was a bit surprised by), if it hadnt i probably would have generated them with AI
