60 lines
2.2 KiB
Plaintext
60 lines
2.2 KiB
Plaintext
# my fork: https://github.com/d-popov/petals-ai/tree/main
|
|
|
|
###get kernel log:
|
|
>journalctl -b
|
|
|
|
### add acpi=off to your kernel boot parameters
|
|
>sudo nano /etc/default/grub
|
|
#Modify the GRUB_CMDLINE_LINUX_DEFAULT line
|
|
++GRUB_CMDLINE_LINUX_DEFAULT="quiet splash acpi=off"
|
|
>sudo update-grub
|
|
|
|
### ignore lid
|
|
>sudo nano /etc/systemd/logind.conf
|
|
+HandleLidSwitch=ignore
|
|
+HandleLidSwitchExternalPower=ignore
|
|
>sudo systemctl restart systemd-logind
|
|
|
|
### install miniconda
|
|
>apt update
|
|
>wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
|
|
>bash Miniconda3-latest-Linux-x86_64.sh
|
|
|
|
> conda create -n petals python=3.10
|
|
> conda activate petals
|
|
> pip install git+https://github.com/bigscience-workshop/petals
|
|
|
|
|
|
|
|
version: '3'
|
|
services:
|
|
dbg-nvidia-smi:
|
|
image: ghcr.io/fnndsc/dbg-nvidia-smi:lite
|
|
|
|
runtime: nvidia
|
|
command: sh -c 'nvidia-smi-wrapper /tmp --cat'
|
|
petals: #https://github.com/bigscience-workshop/petals
|
|
image: learningathome/petals:main
|
|
restart: always
|
|
#command: python -m petals.cli.run_server --port 31330 enoch/llama-65b --adapters timdettmers/guanaco-65b -- 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
|
|
#?meta-llama/Llama-2-70b-chat-hf? meta-llama/Llama-2-70b-hf Enoch/llama-65b-hf ?julep-ai/mpt-30b-orca-mini?
|
|
#command: python -m petals.cli.run_server --use_auth_token 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 --port 31330 --adapters timdettmers/guanaco-65b meta-llama/Llama-2-70b-hf
|
|
# huggingface-cli login [--token TOKEN] [--add-to-git-credential]
|
|
#
|
|
#python -m petals.cli.run_server --port 31330 --adapters timdettmers/guanaco-65b --use_auth_token meta-llama/Llama-2-70b-hf
|
|
command: |
|
|
/bin/bash -c "
|
|
git config --global credential.helper store &&
|
|
huggingface-cli login --token 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 --add-to-git-credential &&
|
|
python -m petals.cli.run_server --port 31330 --adapters timdettmers/guanaco-65b --use_auth_token Enoch/llama-65b-hf
|
|
# Enoch/llama-65b-hf
|
|
# meta-llama/Llama-2-70b-hf
|
|
"
|
|
stdin_open: true
|
|
tty: true
|
|
environment:
|
|
- HUGGINGFACE_TOKEN=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
|
|
ports:
|
|
- 31330:31330
|
|
volumes:
|
|
- /mnt/storage/docker_slow/petals/:/cache |