8-bit MLX quantization of
nvidia/parakeet-tdt-0.6b-v3,
re-hosted byte-for-byte from
animaslabs/parakeet-tdt-0.6b-v3-mlx-8bit
for stable pinning. Conversion chain: NVIDIA base, mlx-community MLX
conversion, animaslabs 8-bit quantization. No architecture, vocabulary, or
training changes anywhere in the chain. See the base model card for
capabilities, training details, and evaluations.
CC-BY-4.0, from the base model, with attribution to NVIDIA (base model),
mlx-community (MLX conversion), and animaslabs (8-bit quantization).
