xTuring VS basaran

Explore the project on GitHub here.

If you're curious, I encourage you to: - Dive deeper with the LLaMA 2 tutorial here. - Explore the project on GitHub here. - Connect with our community on Discord here.

If you want best performance, you need to do both RAG and fine-tuning very well. There are plenty of resources on doing fine-tuning thought. I'm one of the contributors to https://github.com/stochasticai/xturing project focused on fine-tuning LLMs. You can find help in the discord channel listed on the GitHub.

What tools do you use and achieved great results ? … For me i have tried xturing and SFTTrainer and they got me a semi okay results.

Code: https://github.com/stochasticai/xTuring/blob/main/examples/llama/llama_lora_int8.py Colab: https://colab.research.google.com/drive/1SQUXq1AMZPSLD4mk3A3swUIc6Y2dclme?usp=sharing

We are a group of researchers out of Harvard working on open-source library called xTuring, focused on fine-tuning LLMs: https://github.com/stochasticai/xturing.

Fine tuning I was thinking about using this (low GPU memory footprint): https://github.com/stochasticai/xturing/blob/main/examples/int4_finetuning/README.md

I can't speak for QLORA, as I haven't had a chance to get an implementation working, but I've had success with StochasticAI's Xturing. It's by far the most streamlined method of finetuning I've come across, and they offer int8 and int4 fintuning (but only for llama-7B).

What are the current best "no reinventing the wheel" approaches to have Langchain use an LLM through a locally hosted REST API, the likes of Oobabooga or hyperonym/basaran with streaming support for 4-bit GPTQ?

Disclaimer: I am curating LLM-tools on github [1]

A few thoughts:

* allow for custom endpoint URLs, this way people can use open source LLMs with a fake openAI API backend like basaran[2] or llama-api-server[3]

* look into better embedding methods for info-retrieval like InstructorEmbeddings or Document Summary Index

* Don't use a single embedding per content item, use multiple to increase retrieval quality

1 https://github.com/underlines/awesome-marketing-datascience/...

2 https://github.com/hyperonym/basaran

3 https://github.com/iaalm/llama-api-server

open-source alternative to the OpenAI text completion API (https://github.com/hyperonym/basaran)

Guanaco-65B[0] using Basaran[1] for your OpenAI compatible API. You can use any ChatGPT front-end which lets you change the OpenAI endpoint URL.

[0] An fp4 finetune of LLaMA-30B by Tim Dettmers

[1] https://github.com/hyperonym/basaran

For lm-eval, I think you'd either need to take GPTQ's inference script and shim it into a model: https://github.com/EleutherAI/lm-evaluation-harness/tree/master/lm_eval/models or you might be able to use a project like https://github.com/hyperonym/basaran and then you could use the gpt3 model...

There are also: - Basaran repo: "Basaran is an open-source alternative to the OpenAI text completion API. It provides a compatible streaming API for your Hugging Face Transformers-based text generation models". "...Compatibility with OpenAI API and client libraries..."; - llama-cpp-python repo: "Simple Python bindings for @ggerganov's llama.cpp library...". "...OpenAI-like API...".

I want to start with a totally freely available model, so again, that excludes things like LLaMA where the weights are only available through a wait list. The two models that most get my attention and (I think, and hope) fit my criteria of open availability are Cerebras-GPT (13b) and Bloomz (7b). The tools to process and fine-tune that seem most feasible to me, from my limit knowledge, are xturing and basaran.