Ternary-Bonsai-2-27B-DFlash2 — MLX 4-bit

MLX 4-bit affine quantization (group size 64) of the DFlash2 draft head from ProCreations/Ternary-Bonsai-2-27B-DFlash2. Drafts for prism-ml/Ternary-Bonsai-2-27B-mlx-2bit:

mlx_vlm.generate --model prism-ml/Ternary-Bonsai-2-27B-mlx-2bit \
  --draft-model ZeroDegress/Ternary-Bonsai-2-27B-DFlash2-mlx-4bit --draft-kind dflash

file	bytes	sha256
`model.safetensors`	1,082,881,633	`fd9ac742b6c5aa063944c29f54ab3dbee135718d71e5cf8c15e2accfddd9bc7f`
`config.json`	1,322	`16f21748eb15146663898bd4fb9a1dfd0bedefa18632209e7f628b93f394760f`

Weights: Apache-2.0, from ProCreations.

Downloads last month: 667

Safetensors

Model size

2B params

Tensor type

U32

BF16

MLX

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for zerodegress/Ternary-Bonsai-2-27B-DFlash2-mlx-4bit

Base model

Qwen/Qwen3.8-27B

Quantized

prism-ml/Ternary-Bonsai-2-27B-gguf

Quantized

ProCreations/Ternary-Bonsai-2-27B-DFlash2

Quantized

(2)

this model