Qwen3.6-27B-INT8-W8A8
INT8 (W8A8) quantization of Qwen/Qwen3.6-27B — a dense hybrid model (GatedDeltaNet linear-attention + periodic full-attention) with a vision tower and an MTP head.
Quantization
- Symmetric INT8 weights (per-channel, MSE observer) + INT8 dynamic per-token activations (llm-compressor). int8 is robust, so no AWQ/SmoothQuant smoothing is needed.
- Quantized: all MLP, the full-attention
self_attn, and the GatedDeltaNetout_proj. - Kept bf16 (quality-sensitive): GatedDeltaNet
in_proj_a/in_proj_b,lm_head, embeddings, vision tower, MTP head. - Format:
compressed-tensors(int-quantized). Full recipe:recipe.yaml.
INT8 W8A8 is near-lossless; for the smallest footprint use the INT4-W4A16 variant (GSM8K 96.8% / MMLU-Pro 82.4%).
Usage (vLLM)
vllm serve Avesed/Qwen3.6-27B-INT8-W8A8 \
--tensor-parallel-size 2 --trust-remote-code --reasoning-parser qwen3
Served via vLLM's INT8 W8A8 CUTLASS path (works on Ampere sm_80 / sm_86).
Quantized with vllm-ampere-optimized/quantize.