catalyst
Catalyst. Accelerated deep learning R&D with PyTorch.
Does not declare an LLM SDK itself, but depends on a package that does, 1 step away.
Advisories
Advisories that name catalyst as affected. Advisory records do not state an ecosystem, so packages with the same name in other ecosystems also match. For a version-exact check of your own dependencies, use Stack Check.
No advisory in the database names this package.
How it reaches an LLM package
Shortest path through the runtime dependencies of each latest release.
- catalyst
- accelerate
Dependencies of the latest release
As declared in PyPI metadata for version 22.4. Optional extras are listed with their extra name.
- accelerateHugging Face Hub / Transformers
- deepspeed[deepspeed]Hugging Face Hub / Transformers
- mlflow[mlflow]FastMCP, LangChain
- wandb[wandb]Google Vertex AI SDK
- comet-ml[comet]reaches LLM
- black[dev]
- catalyst-codestyle[dev]
- catalyst-sphinx-theme[dev]
- click[dev]
- docutils[dev]
- hydra-slayer
- imageio[all]
- jinja2[dev]
- matplotlib[all]
- mock[dev]
- neptune-client[neptune]
- numpy
- onnx[onnx]
- onnxruntime[onnx]
- onnxruntime-gpu[onnx-gpu]
- opencv-python-headless[all]
- optuna[all]
- pandas[all]
- path[dev]
- pillow[all]
- pre-commit[dev]
- pytest[dev]
- requests[all]
- scikit-image[all]
- scikit-learn[all]
- scipy[all]
- sphinx[dev]
- tensorboardx
- tomlkit[dev]
- torch
- torch-tb-profiler[profiler]
- torchvision[all]
- tqdm
Tracked packages that depend on it
Among the packages in the registry; not every dependent on PyPI.