Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.maint.python > #15546

Bug#1065572: ITP: llm -- Access large language models from the command-line

Cross-posted to 3 groups.

Show all headers | View raw

Package: wnpp
Severity: wishlist
Owner: Antoine Beaupre <anarcat@debian.org>
X-Debbugs-Cc: debian-devel@lists.debian.org, debian-python@lists.debian.org, debian-ai@lists.debian.org

* Package name    : llm
  Version         : 0.13.1
  Upstream Contact: Simon Willison <swillison@gmail.com>
* URL             : https://llm.datasette.io/
* License         : Apache-2.0
  Programming Lang: Python
  Description     : CLI utility and Python library for interacting
  with Large Language Models

A CLI utility and Python library for interacting with Large Language
Models, both via remote APIs and models that can be installed and run
on your own machine.

Run prompts from the command-line, store the results in SQLite,
generate embeddings and more.

----

So there has been some discussions about packaging LLM things in
Debian, and this is my stab at opening a discussion about *what*
exactly, we *can* package, right now.

LLM is a (possibly poorly named) package that allows users to interact
with various language models. Its primary target is obviously the
OpenAI API (it's the example in the "Quick Start"), but it also
supports local models like Llama 2, Ollama, and MLC, and other online
models like Claude, GEmini and Mistral.

I *think* this belongs in contrib, because it *mostly* relies on
non-free software (and services) to do its thing, but I'd be happy to
move that around.

I need this because I was using GPT plus for a while and now I'm
switching to the API to cut down on costs.

Back to linux.debian.maint.python | Previous | Next | Find similar | Unroll thread

Thread

Bug#1065572: ITP: llm -- Access large language models from the command-line Antoine Beaupre <anarcat@debian.org> - 2024-03-06 21:30 +0100

csiph-web