2023-03-28 13:24:05 -07:00
2023-03-25 16:23:14 +00:00
2023-03-27 16:33:00 +00:00
2023-03-25 16:23:14 +00:00
2023-03-27 17:30:33 +00:00
2023-03-27 17:30:44 +00:00
2023-03-25 16:16:40 +00:00
2023-03-28 20:22:48 +00:00
2023-03-28 18:47:38 +00:00
2023-03-28 00:09:47 +00:00
GIF
2023-03-28 15:54:44 -04:00
2023-03-25 16:17:48 +00:00
2023-03-28 16:23:16 -04:00
2023-03-25 16:16:40 +00:00
2023-03-27 17:30:24 +00:00
2023-03-28 13:24:05 -07:00

GPT4All

Demo, data and code to train an assistant-style large language model

📗 Technical Report

gpt4all-lora-demo

Try it yourself

You can download pre-compiled LLaMa C++ Interactive Chat binaries here:

and the model

Place the binary and quantized model in the same directory and start chatting!

To compile for custom hardware, see our fork of the Alpaca C++ repo.

Reproducibility

Trained LoRa Weights:

Raw Data:

We are not distributing a LLaMa 7B checkpoint.

You can reproduce our trained model by doing the following:

Setup

Clone the repo

git clone --recurse-submodules git@github.com:nomic-ai/gpt4all.git

git submodule configure && git submodule update

Setup the environment

python -m pip install -r requirements.txt

cd transformers
pip install -e . 

cd ../peft
pip install -e .

Training

accelerate launch --dynamo_backend=inductor --num_processes=8 --num_machines=1 --machine_rank=0 --deepspeed_multinode_launcher standard --mixed_precision=bf16  --use_deepspeed --deepspeed_config_file=configs/deepspeed/ds_config.json train.py --config configs/train/finetune-7b.yaml

Generate

python generate.py --config configs/generate/generate.yaml --prompt "Write a script to reverse a string in Python

If you utilize this reposistory, models or data in a downstream project, please consider citing it with:

@misc{gpt4all,
  author = {Yuvanesh Anand and Zach Nussbaum and Brandon Duderstadt and Benjamin Schmidt and Andriy Mulyar},
  title = {GPT4All: Training an Assistant-style Chatbot with Large Scale Data Distillation from GPT-3.5-Turbo},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/nomic-ai/gpt4all}},
}
Description
gpt4all: an ecosystem of open-source chatbots trained on a massive collections of clean assistant data including code, stories and dialogue
Readme 63 MiB
Languages
C++ 51.9%
QML 30.3%
Python 7.6%
CMake 5.4%
JavaScript 3.2%
Other 1.4%