Python SDK, command-line tools, runnable recipes, and documentation for Cortex Training through the Cortex Training SNOWAPI.
- Follow the getting-started path
- Run the first supervised fine-tuning job
- Browse runnable recipes
- Check model and training-method compatibility
- Use the CLI and Python client
- Read the REST API reference
Requires Python 3.8 or later.
pip install git+https://github.com/Snowflake-AI-Research/cortex-training.gitFor local development:
git clone https://github.com/Snowflake-AI-Research/cortex-training.git
cd cortex-training
pip install -e .The package installs:
cortex-training, for submitting and managing jobscortex-training tui, for viewing job logscortex_training, the Python SDK
Verify the command entry points:
cortex-training --help
cortex-training tui --helpcortex-training list
cortex-training submit examples/api/training.json
cortex-training tui JOB_IDfrom cortex_training import CortexTrainingClient, CortexTrainingEngineConnection settings use CORTEX_TRAINING_* and SNOWFLAKE_* environment
variables. Login state is stored under ~/.config/cortex-training/, and TUI
cache state is stored under ~/.cache/cortex-training/, unless their existing
override variables are used.
See the CLI reference for commands and configuration.
| Path | Purpose |
|---|---|
docs/ |
Getting started material, concepts, guides, and reference |
recipes/ |
End-to-end training, sampling, and evaluation workflows |
examples/api/ |
Small JSON examples for individual API operations |
examples/config/ |
Connection configuration templates |
src/cortex_training/ |
Installable Python client |
tests/ |
Client and CLI tests |
The current onboarding work is tracked in docs/internal/onboarding-roadmap.md.