Downloads · 30 days
60
24% of all-time downloads
shoemoney/Qwen3.8-27B-Abliterated-MLX-q5
Qwen3.8-27B-Abliterated-MLX-q5 is a machine learning model from shoemoney. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for mlx-vlm. The card lists the license as apache-2.0.
MLX 5-bit quantisation of huihui-ai/Huihui-Qwen3.8-27B-abliterated.
Downloads · 30 days
60
24% of all-time downloads
All-time downloads
252
Public
Parameters
27.4B
19.4 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors19.4 GB · 100%
How the weights are stored.
U3226.9B · 98%
From the Hugging Face model README
MLX 5-bit quantisation of huihui-ai/Huihui-Qwen3.8-27B-abliterated.
Changes: weights quantised to 5-bit from the BF16 source with
mlx_vlm.convert. No fine-tuning, no merging, no re-alignment.
Converted and measured on one machine — Apple M3 Ultra, 96 GB unified memory, macOS 27 — as part of a full ladder. Every rung in this family came from the same BF16 source with the same group size, so bit width is the only variable between them.
| Size on disk | 19.44 GB |
| Perplexity | 6.419 |
| Relative to best rung in family | 1.03× |
| Throughput (1 req / 8 concurrent) | 22.1 / 65.7 tok/s |
Perplexity measured on allenai/tulu-3-sft-mixture, 192 samples of 512 tokens,
seed 123 — identical for every rung.
Perplexity is only comparable within this family. Tokenizers differ between model families, so a number here should never be compared against a different base model's. The
×column above is the meaningful one.
pip install mlx-vlm
mlx_vlm.generate --model shoemoney/Qwen3.8-27B-Abliterated-MLX-q5 --prompt "Hello" --max-tokens 256
Load with mlx-vlm, not mlx-lm — this architecture is registered in mlx-vlm.
mlx_vlm.convert --hf-path huihui-ai/Huihui-Qwen3.8-27B-abliterated \
--mlx-path Qwen3.8-27B-Abliterated-q5 -q --q-bits 5 --q-group-size 64
apache-2.0, inherited from the base model. Attribution above.