nvidia-tao-daft 7.0.0

NVIDIA TAO Dataset Annotation Format Toolkit - A comprehensive toolkit for organizing, validating, and loading annotated datasets for computer vision and vision-language models

NVIDIA TAO DAFT: Dataset Annotation Format Toolkit

A toolkit for vision-language dataset formats — JSON-schema specs plus CLI/Python tools to validate and convert between them.

Overview

VLM dataset workflows have a contract-drift problem. Annotation pipelines emit data in one shape, training pipelines expect another, and the glue between them is ad-hoc adapter code that silently goes stale. Field renames, new optional values, schema-vs-data mismatches — these surface as training-time bugs rather than at the producer / consumer boundary where they belong.

What DAFT is:

• Schemas for vision-language dataset shapes — both annotation (what producers emit) and training (what consumers expect).
• A CLI + validator so anyone holding a dataset can check it against its schema before handing it off.
• Converters between annotation and training shapes — explicit, deterministic, with optional flags for media handling.
• A reference Python adapter that plugs one of the training shapes into cosmos-rl SFT.

New formats, validators, converters, and adapters are welcome; the same registration pattern that wires the built-ins works for your own extensions.

Value, by audience:

For…	DAFT gives you…
Producers (annotation pipelines, human annotators)	Target one of these schemas and your output is consumable by any downstream tool that speaks the same schema.
Consumers (training pipelines, researchers)	Validate your input dataset before launching a training run. If it passes, your loader contract holds.

Quick start

# Install (direct from git)
pip install git+https://gitlab-master.nvidia.com/nvidia-tao-toolkit/experimental/nvidia-tao-daft.git

# Install (from wheel)
pip install nvidia-tao-daft

# Verify
tao-daft --help

For runnable examples, see examples/ and the CLI reference.

Documentation

Area	What's there	Link
Formats	Format registry, per-format specs (metropolis-v3.0, cosmos-reason-v1.0, tao-vl-reason-v1.0), versioning policy	formats
CLI	tao-daft validate / convert reference	cli
Validators	Validation engine	validators
Converters	Conversion pairs and pair-specific options	converters
Datasets	Training-loop adapters (cosmos-rl)	datasets
Examples	Working datasets per format	examples

Repository structure

nvidia-tao-daft/
├── examples/datasets/        # Working datasets, one subdir per format
│
├── tests/                    # Test suite (schemas, validators, converters, CLI, doc consistency)
│
└── src/nvidia_tao_daft/
    ├── cli/                  # tao-daft entry point (validate, convert)
    ├── formats/              # Format specifications + JSON schemas
    ├── validators/           # Validation engine
    ├── converters/           # Format converters (pairs/)
    └── datasets/             # Training-loop adapters

Requirements

Python 3.10 – 3.13. Runtime dependencies: jsonschema, pydantic. Dev dependencies: see pyproject.toml.

Contributing

See CONTRIBUTING.md for the DCO sign-off requirement.

License

Apache 2.0.