SofiTesfay2010/Qwen3.8-27B-Uncensored-Compressed

1.25 bits per weight (Sherry 3:4) compressed version of orcarouter/Qwen3.8-27B-Uncensored.

Specs

  • Method: Sherry 3:4 Sparse Ternary Quantization ($C(4,3) \times 2^3 = 32 = 2^5$)
  • Size: ~4.3 GB (down from ~55.6 GB)
  • Memory Footprint: Runs comfortably in <5 GB RAM / VRAM
Downloads last month
-
Safetensors
Model size
7B params
Tensor type
I32
F16
U8
F8_E4M3
Inference Providers NEW
This model isn't deployed by any Inference Provider. 馃檵 Ask for provider support

Model tree for SofiTesfay2010/Qwen3.8-27B-Uncensored-Compressed

Base model

Qwen/Qwen3.8-27B
Finetuned
(6)
this model