The Local LLM Index / Quantization & Formats / #100

woct0rdho/transformers-qwen3-moe-fused

by woct0rdho · Quantization & Formats · updated 3d ago

Fused Qwen3 MoE layer for faster training, compatible with Transformers, LoRA, bnb 4-bit quant, Unsloth. Also possible to train LoRA over GGUF

59
momentum
260
stars
18
forks
#100
rank
View on GitHub →