sedpack
Creates, writes, inspects, and iterates multi-shard machine learning and side-channel analysis datasets using the sedpack library (sedpack.io). Use when creating a sedpack Dataset, defining Metadata, DatasetStructure, or Attribute schemas, writing examples with dataset.filler(), DatasetFiller, or dataset.write_multiprocessing(), iterating data for JAX, PyTorch, Keras, TensorFlow, or NumPy (as_numpy_iterator, as_numpy_iterator_concurrent, as_numpy_iterator_async, as_numpy_iterator_rust, as_numpy_iterator_rust_batched, as_tfdataset), filtering or balancing shards via custom_metadata, or converting raw datasets (HDF5, NPZ, traces) into sedpack format. Don't use for modifying sedpack's internal Rust or FlatBuffers compiler implementation unless high-level API semantics are needed.
Source facts
- Repository
- google/sedpack
- Last source activity
- November 9, 2026 at 10:22
- Detected SKILL.md language
- English
- Stars
- 39
- Forks
- 14
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Showing SKILL.md
| name | sedpack |
|---|---|
| description | Creates, writes, inspects, and iterates multi-shard machine learning and side-channel analysis datasets using the sedpack library (sedpack.io). Use when creating a sedpack Dataset, defining Metadata, DatasetStructure, or Attribute schemas, writing examples with dataset.filler(), DatasetFiller, or dataset.write_multiprocessing(), iterating data for JAX, PyTorch, Keras, TensorFlow, or NumPy (as_numpy_iterator, as_numpy_iterator_concurrent, as_numpy_iterator_async, as_numpy_iterator_rust, as_numpy_iterator_rust_batched, as_tfdataset), filtering or balancing shards via custom_metadata, or converting raw datasets (HDF5, NPZ, traces) into sedpack format. Don't use for modifying sedpack's internal Rust or FlatBuffers compiler implementation unless high-level API semantics are needed. |