Ir para o conteúdo
ModelListadosExterno
Baixar

Huihui ThinkingCap Qwen3.6 27B Abliterated MLX 4bit MTP

MTPLX-compatible conversion of huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated, pinned to revision 44f63da.

Fontehuggingface.co/pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP/tree/e44265b8a62105a775011e5be92099537a5ee671
TipoText modelModelo basehuihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliteratedVersãoBF16Bnota

Fonte do modelo

Tipo
Text model
Tamanho do arquivo
810 MB
Modelo base
huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated
Versão
vBF16
Atualizado
30 de out. de 2026
Downloads
144
Fonte
Hugging Face

Descrição da fonte

Huihui ThinkingCap Qwen3.6 27B Abliterated — MLX 4-bit MTP

MTPLX-compatible conversion of huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated, pinned to revision 44f63da.

Runtime-specific artifact: do not load this repository in LM Studio. LM Studio flattens the MTPLX sidecar into its target directory, causing the target loader to reject the 15 mtp.* tensors. For the recommended MTP experience, use the oMLX Native-MTP model; oMLX is the faster, more mature integrated path for this model.

  • Trunk: MLX affine 4-bit, group size 64.
  • MTP head: 15-tensor BF16 sidecar at mtp/weights.safetensors.
  • Runtime: MTPLX 2.1.0, Apple Silicon only.
pip install -U mtplx
mtplx run --model pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP --depth 3 "Hello"

MTPLX Forge classified this artifact as verified-native. A bounded 32-token M4 Max verification measured 24.94 tok/s AR and 46.02 tok/s at MTP depth 3 (1.85x); treat this as a packaging smoke, not a general benchmark.

This is an abliterated model with reduced safety behavior. Review the upstream model card before use.

Fontes

1 fonte
Hugging FacePrimárioVersão e44265b8a62105a775011e5be92099537a5ee671
huggingface.co/pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP/tree/e44265b8a62105a775011e5be92099537a5ee671

Verificado 30 de out.

Download files

Model · 3 weight files

All files listed in the provider index are present.

Repository files

File inventory only; completeness is unverified.

Artefatos de modelo

1 artefato

mtp/weights.safetensors

safetensors · BF16 (tensor header) · 810 MB · SHA-256 86671263cf45…d264 · Hugging Face

Download file

Trechos de fonte

3 trechos
Readme14 de ago.huggingface.co/pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP

--- license: apache-2.0 base_model: huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated base_model_relation: quantized library_name: mlx pipeline_tag: image-text-to-text tags: - mlx - mtplx - mtp - qwen3_6 - 4-bit - abliterated --- # Huihui ThinkingCap Qwen3.6 27B Abliterated — MLX 4-bit MTP MTPLX-compatible conversion of [huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated](https://huggingface.co/huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated), pinned to [revision `44f63da`](https://huggingface.co/huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated/commit/44f63da8141407af529405c1e4b83fa39b70abe0). > **Runtime-specific artifact:** do not load this repository in LM Studio. LM Studio flattens the MTPLX sidecar into its target directory, causing the target loader to reject the 15 `mtp.*` tensors. For the recommended MTP experience, use the [oMLX Native-MTP model](https://huggingface.co/pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-oMLX-MTP); oMLX is the faster, more mature integrated path for this model. - Trunk: MLX affine 4-bit, group size 64. - MTP head: 15-tensor BF16 sidecar at `mtp/weights.safetensors`. - Runtime: [MTPLX 2.1.0](https://github.com/youssofal/MTPLX), Apple Silicon only. ```bash pip install -U mtplx mtplx run --model pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP --depth 3 "Hello" ``` MTPLX Forge classified this artifact as `verified-native`. A bounded 32-token M4 Max verification measured 24.94 tok/s AR and 46.02 tok/s at MTP depth 3 (1.85x); treat this as a packaging smoke, not a general benchmark. This is an abliterated model with reduced safety behavior. Review the [upstream model card](https://huggingface.co/huihui-ai/Huihui-ThinkingCap-Qwen3.6-27B-abliterated) before use.

Note30 de out.huggingface.co/pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP

Source context: 144 downloads · 1 likes · Pipeline image-text-to-text · Library mlx · Repo pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP

Note14 de ago.huggingface.co/pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP

Source context: 865 downloads · 1 likes · Pipeline image-text-to-text · Library mlx · Repo pixelkaiser/Huihui-ThinkingCap-Qwen3.6-27B-abliterated-MLX-4bit-MTP