Chat test scripts for LLM
Git clone
$ cd && git clone https://github.com/karaage0703/ChatLLM
Setup devcontainer
CPU
$ docker compose up base
GPU
$ docker compose up gpu
GPU(Nvidia container)
$ docker compose up nvidia
Check docker image name
$ docker ps
$ docker exec -it <image name> /bin/bash
# example
# $ docker exec -it chatllm-nvidia-1 /bin/bash
root@hostname:/# cd /root
root@hostname:~# python3 chat_calm.py
root@hostname:~# python3 chat_rinna.py
root@hostname:~# python3 chat_rwkv.py
root@hostname:~# python3 chat_llama2.py
root@hostname:~# python3 chat_weblab.py
root@hostname:~# python3 chat_elyza.py
$ cd ~/ChatLLM
$ docker build -t ubuntu:ChatLLM .
Use GPU
$ cd ~/ChatLLM
$ docker run -it --rm -v $(pwd):/root --gpus all ubuntu:ChatLLM
Use CPU
$ cd ~/ChatLLM
$ docker run -it --rm -v $(pwd):/root ubuntu:ChatLLM
root@hostname:~# python3 chat_calm.py
root@hostname:~# python3 chat_rinna.py
root@hostname:~# python3 chat_rwkv.py
root@hostname:~# python3 chat_llama2.py
for stablelm
root@hostname:/# huggingface-cli login
root@hostname:~# python3 chat_stablelm.py
- https://huggingface.co/cyberagent/open-calm-7b
- https://huggingface.co/rinna/japanese-gpt-neox-3.6b-instruction-sft
- https://nowokay.hatenablog.com/entry/2023/05/17/144518
- https://note.com/npaka/n/n401dccfadedc
- https://note.com/npaka/n/n0ad63134fbe2
- https://zenn.dev/karaage0703/articles/2b753b4dc26471
- https://zenn.dev/karaage0703/articles/d58d79d8e77ab8
- https://zenn.dev/karaage0703/articles/d3893b551c68fa
- https://note.com/npaka/n/nfacbeb1ae709
- https://colab.research.google.com/github/mkshing/notebooks/blob/main/stabilityai_japanese_stablelm_alpha_7b.ipynb
- https://gitlab.com/nvidia/container-images/cuda/blob/master/doc/supported-tags.md
- https://note.com/npaka/n/nd05b0334cb75
- https://zenn.dev/tsuzukia/articles/f886a93fc1f2fa
- https://note.com/npaka/n/nbb94b45f47a5