---
license: apache-2.0
base_model:
- Qwen/Qwen3-VL-30B-A3B-Instruct
---

# Quantized versions of Qwen3-VL-30B-A3B-Instruct with mmproj

my motivation for this was that most/all other quants i seen of this is they only have an .mmproj file for f16,</br>
but this file needs to be of same quant as the model itself</br>
(locking you to only using f16 if you want the vision support, which is ofcourse the whole point of the model)</br>

NOTE: im inexperienced in quantizing, its possible i have done mistakes or missed something, possibly reducing or degrading the quality of this model