parakeet-tdt-0.6b-v3 - Core ML

Core ML build of nvidia/parakeet-tdt-0.6b-v3 (FastConformer encoder + TDT decoder, 0.6B parameters, 25 European languages including Russian and Ukrainian, punctuation and capitalization built in).

Mirrored from FluidInference/parakeet-tdt-0.6b-v3-coreml for the Letopis on-device transcription pipeline; the runtime is the FluidAudio Swift package (Apache-2.0). This mirror carries the single runtime file set the app downloads (fp16/int8 inference path).

Files

  • Preprocessor.mlmodelc, Encoder.mlmodelc, Decoder.mlmodelc, JointDecisionv3.mlmodelc
  • config.json, parakeet_vocab.json, parakeet_v3_vocab.json

Credits and license

  • Model: NVIDIA parakeet-tdt-0.6b-v3, CC-BY-4.0.
  • Core ML conversion: FluidInference.
  • No changes to the weights; this is a byte-identical mirror of the runtime files.
Downloads last month
23
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for smkrv/parakeet-tdt-0.6b-v3-coreml

Finetuned
(62)
this model