Parakeet TDT-CTC 0.6B Japanese — Core ML (Diction mirror)
The Core ML build that the Diction iOS app downloads for its on-device
Japanese dictation model. It is a byte-identical copy of the files the app loads —
Preprocessor.mlmodelc, Encoder.mlmodelc, Decoderv2.mlmodelc, Jointerv2.mlmodelc and the
root JSON files — from
FluidInference/parakeet-0.6b-ja-coreml
(revision 2952296ff1da4a6d6a7aec545e226367db80c612), mirrored so the app downloads from a
namespace we control. Load it with FluidAudio's AsrModels (version .tdtJa).
License & attribution
The weights are NVIDIA's parakeet-tdt_ctc-0.6b-ja, released by NVIDIA under CC-BY-4.0. The Core ML conversion is by FluidInference, under the same licence. This mirror is redistributed under CC-BY-4.0 (copy in LICENSE); no changes were made to the model files. Retain this attribution to NVIDIA and FluidInference when redistributing.
- Downloads last month
- 43
Model tree for DictionLabs/parakeet-0.6b-ja-coreml
Base model
nvidia/parakeet-tdt_ctc-0.6b-ja