shishirpatil

6 ms·

1. 31.▲
   by shishirpatil 3y ago

   Haha, was busy yesterday! Or was I? :P
2. 32.▲
   by shishirpatil 3y ago

   Thanks @sfriedr We generate self-instruct data and then fine tune the base model with perplexity loss. The self-instruct data is https://github.com/ShishirPatil/gorilla/tree/main/data/apibe... Thank
3. 33.▲
   by shishirpatil 3y ago

   Yes, when we released the initial set of models Open Llama was still at 600B checkpoint and not finished training yet. It's an easy port :)
4. 34.▲
   by shishirpatil 3y ago

   Thanks for the kind words @gsharma! Private APIs is something we are definitely thinking about.
5. 35.▲
   by shishirpatil 3y ago

   Hi swyx, the weights are also open sourced at https://huggingface.co/gorilla-llm Let us know if you are unable to access them.
6. 36.▲
   by shishirpatil 3y ago

   Hi @csjh, we trained to model to also additionally output additional context so it would be useful for a downstream task. We wrapped the API call with special decorator so it's easier to just regex. Would you like to just have the API
7. 37.▲
   by shishirpatil 3y ago

   Yes we have three set of models. One based on llama - which you are right, cannot be used commercial. We have two additional models based on MPT-7 base and Falcon-7B which can be used commercially with no obligations!
8. 38.▲
   by shishirpatil 3y ago

   Good point. This was the original release, we now also have Apache-2.0 licensed models finetuned on MPT-7B and Falcon-7B!
9. 39.▲
   by shishirpatil 3y ago

   Langchain is a terrific project that tries to teach agents how to use tools using prompting. Our take on this is that prompting is not scalable if you want to pick between 1000s of APIs. So Gorilla is a LLM that can pick and write the seman
10. 40.▲
    by shishirpatil 3y ago

    Hi HN, I'm one of the lead authors on the paper! Gorilla is an open source effort and we would love to hear from the community. Let us know if you have any questions or suggestions!!
11. 41.▲
    Gorilla LLM is now Apache 2.0 – can be used commercially(github.com)

    9 points by shishirpatil 3y ago|1 comments
12. 42.▲
    Gorilla: Large Language Model connected with massive APIs(gorilla.cs.berkeley.edu)

    255 points by shishirpatil 3y ago|119 comments
13. 43.▲
    by shishirpatil 4y ago

    A good 32% reduction taking it to less than a 1$ per TB.