Hy-MT2-1.8B, 4-bit MLX build
A 4-bit quantization of tencent/Hy-MT2-1.8B, Tencent's translation model for 33 languages, converted for Apple silicon with mlx-lm:
mlx_lm convert --hf-path tencent/Hy-MT2-1.8B -q --q-bits 4
Weights are unchanged apart from the quantization (4.5 bits per weight, group size 64). Built for Awnsy, a macOS menu bar translator, where it runs on-device.
Licensed under Apache 2.0, as the original.
- Downloads last month
- 148
Model size
2B params
Tensor type
U32
·
BF16 ·
Hardware compatibility
Log In to add your hardware
4-bit
Model tree for aidsoid/Hy-MT2-1.8B-MLX-4bit
Base model
tencent/Hy-MT2-1.8B