Quickstart¶
Prerequisites
To use the open-source version, make sure to install the server and configure backends.
If you're using dstack Sky , install the CLI and run the dstack config command:
Once the CLI is set up, follow the quickstart.
Initialize a repo¶
To use dstack's CLI in a folder, first run dstack init within that folder.
$ mkdir quickstart && cd quickstart
$ dstack init
Your folder can be a regular local folder or a Git repo.
Define a configuration¶
Define what you want to run as a YAML file. The filename must end with .dstack.yml (e.g., .dstack.yml
or train.dstack.yml are both acceptable).
Dev environments allow you to quickly provision a machine with a pre-configured environment, resources, IDE, code, etc.
type: dev-environment
# Use either `python` or `image` to configure environment
python: "3.11"
# image: ghcr.io/huggingface/text-generation-inference:latest
ide: vscode
# (Optional) Configure `gpu`, `memory`, `disk`, etc
resources:
gpu: 80GB
Tasks make it very easy to run any scripts, be it for training, data processing, or web apps. They allow you to pre-configure the environment, resources, code, etc.
type: task
python: "3.11"
env:
- HF_HUB_ENABLE_HF_TRANSFER=1
commands:
- pip install -r fine-tuning/qlora/requirements.txt
- python fine-tuning/qlora/train.py
# (Optional) Configure `gpu`, `memory`, `disk`, etc
resources:
gpu: 80GB
Ensure requirements.txt and train.py are in your folder. You can take them from dstack-examples.
Services make it easy to deploy models and apps cost-effectively as public endpoints, allowing you to use any frameworks.
type: service
image: ghcr.io/huggingface/text-generation-inference:latest
env:
- MODEL_ID=mistralai/Mistral-7B-Instruct-v0.1
commands:
- text-generation-launcher --port 8000 --trust-remote-code
port: 8000
# (Optional) Configure `gpu`, `memory`, `disk`, etc
resources:
gpu: 80GB
Run configuration¶
Run a configuration using the dstack run command, followed by the working directory path (e.g., .), the path to the
configuration file, and run options (e.g., configuring hardware resources, spot policy, etc.)
$ dstack run . -f train.dstack.yml
BACKEND REGION RESOURCES SPOT PRICE
tensordock unitedkingdom 10xCPU, 80GB, 1xA100 (80GB) no $1.595
azure westus3 24xCPU, 220GB, 1xA100 (80GB) no $3.673
azure westus2 24xCPU, 220GB, 1xA100 (80GB) no $3.673
Continue? [y/n]: y
Provisioning...
---> 100%
Epoch 0: 100% 1719/1719 [00:18<00:00, 92.32it/s, loss=0.0981, acc=0.969]
Epoch 1: 100% 1719/1719 [00:18<00:00, 92.32it/s, loss=0.0981, acc=0.969]
Epoch 2: 100% 1719/1719 [00:18<00:00, 92.32it/s, loss=0.0981, acc=0.969]
The dstack run command automatically uploads your code, including any local uncommitted changes.
To exclude any files from uploading, use .gitignore.