W4A16 quantization using llmcompressor. Run with:

vllm serve leon-se/gemma-3-27b-it-qat-W4A16-G128 --max-model-len 4096 --max-num-seqs 1

Safetensors

Model size

6.64B params

Tensor type

I64

I32

BF16

Inference Providers NEW

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

Model tree for leon-se/gemma-3-27b-it-qat-W4A16-G128

Base model

Finetuned

Finetuned

Quantized

(10)

this model