Quantization format

TQ2_0

TQ2_0 has a nominal rate of bits per weight. We have no measured files carrying this label yet.

From the file· 19 files measured
Nominal bpw
from the block layout
Measured average
2.808
19 files
Range
2.069–4.495
varies by architecture
File sizes
0.02 GiB+
up to 8.82 GiB

Real files

one per model, most downloaded first
ModelSizeEffective bpwvs nominal
Qwen-Image-Edit-25096.52 GiB2.739
Qwen3-0.6B-Base0.31 GiB4.495
Qwen-Image-Edit6.55 GiB2.753
HopCoder-Mini-35B-A3B-VL368.82 GiB2.158
Wan_2.2_ComfyUI_Repackaged1.52 GiB
gemma-4-26B-A4B8.71 GiB2.819
Llama-3.2-3B0.99 GiB2.635
diffusiongemma-26B-A4B-it-HERETIC-Uncensored8.72 GiB2.900
Qwen3.5-9B-Base2.99 GiB2.656
Qwen3.5-2B0.75 GiB2.814
granite-3.1-8b-instruct2.07 GiB2.176
Supra-Router-51M0.02 GiB4.077
Maestro1-9B2.48 GiB2.426
bitnet_b1_58-large0.20 GiB2.383
BugTraceAI-Apex-G4-26B-Master-f168.71 GiB2.965
Supra-50M-Base0.02 GiB4.077
Ternary-Bonsai-8B-gguf1.97 GiB2.069
granite-3.1-2b-instruct0.66 GiB2.249
Ternary-Bonsai-8B2.25 GiB2.359
From the filewhat these mean

Other quantizations