Edit on GitHub
  • John Snow Labs
  • Mar 29, 2025

Swahili (macrolanguage) wav2vec2_large_xls_r_300m_swahili_pipeline pipeline Wav2Vec2ForCTC from thinkKenya

  • sw
  • open_source
  • pipeline
  • onnx

Description

Pretrained Wav2Vec2ForCTC, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP.wav2vec2_large_xls_r_300m_swahili_pipeline is a Swahili (macrolanguage) model originally trained by thinkKenya.

Download Copy S3 URI

How to use


pipeline = PretrainedPipeline("wav2vec2_large_xls_r_300m_swahili_pipeline", lang = "sw")
annotations =  pipeline.transform(df)   



val pipeline = new PretrainedPipeline("wav2vec2_large_xls_r_300m_swahili_pipeline", lang = "sw")
val annotations = pipeline.transform(df)

Model Information

Model Name: wav2vec2_large_xls_r_300m_swahili_pipeline
Type: pipeline
Compatibility: Spark NLP 5.5.1+
License: Open Source
Edition: Official
Language: sw
Size: 1.2 GB

References

https://huggingface.co/thinkKenya/wav2vec2-large-xls-r-300m-sw

Included Models

  • AudioAssembler
  • Wav2Vec2ForCTC