pip install scattermoe@git+https://github.com/shawntan/scattermoe.git@63b76a2f5f28c052fb4cd7c34479a54158354052

To convert back to standard Qwen3Moe look at convert_back.py

Safetensors

Model size

30.5B params

Tensor type

BF16

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for chargoddard/Qwen3-30B-A3B-Base-ScatterMoE

Base model

Finetuned

Finetuned

(30)

this model