WangChanGLM API

API Service for the WangChanGLM model. More information about the model can be found here.

Pairs nicely with the WangChanGLM App.

Try out the public demo at https://wangchanglm.in.th

Build docker

docker build -t scb10x/thaillm .

Run docker

mkdir .cache #optional
chmod 777 .cache #optional

# on cpu
docker run --restart unless-stopped -p 80:7860 -v `pwd`/.cache:/home/user/.cache -dt scb10x/thaillm

# on gpu
docker run --gpus all --restart unless-stopped -p 80:7860 -v `pwd`/.cache:/home/user/.cache -dt scb10x/thaillm

Request with cURL

curl --request POST \
  --url http://localhost:80/generate \
  --header 'Content-Type: application/json' \
  --data '{"instruction": "เขียนบทความเกี่ยวกับ ประโยชน์ของการออกกำลังกาย"}'

Deploy on GCP

Create a VM

1. Compute Engine > Create an instance
2. GPUs > NVIDIA T4
3. Machine Type > n1-standard-8
4. Boot disk > Switch Image > Debian 10 based Deep Learning VM with M108 > 250GB
5. Firewall > Allow HTTP traffic > Allow HTTPS traffic
6. Create

Install git-lfs on the VM

curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | sudo bash
sudo apt-get install git-lfs

SSH into the VM and build

1. clone this repo
2. build docker: ./scripts/build.sh
3. run docker: ./scripts/start.sh