Sign Language Processing: TFDS Loaders for Sign Language Research

The sign-language-processing organization provides this collection of TensorFlow Datasets (TFDS) loaders for various sign language datasets as of February 2026. It standardizes the ingestion of sign language video data and associated annotations for use in machine learning frameworks.

Use Cases

• Training sign language recognition models using video frame sequences and gloss annotations
• Developing sign-to-text translation systems using parallel video and natural language text pairs
• Benchmarking sign language production models using skeletal or pose data provided by specific loaders

Strengths

• Standardized TFDS format for seamless integration with TensorFlow data pipelines
• Maintained by the sign-language-processing organization, a specialized research group
• Updated as of February 2026, ensuring compatibility with recent software versions

Limitations

• Requires a TensorFlow environment, which may limit users of other frameworks like PyTorch
• Metadata does not specify the total number of records or hours across the included loaders
• Dependency on external dataset hosting; if the original source is offline, the loader may fail

Provenance

Source

sign-language-processing organization on GitHub

Collection Method

Aggregation and formatting of existing sign language datasets into the TFDS format

Freshness

Updated as of February 2026.