[Docs] Use trtllm_mha for Qwen3.6 B300 (#29964)

This commit is contained in:
Mohammad Miadh Angkad
2026-07-07 01:44:01 -07:00
committed by GitHub
parent 9ddea8d9ef
commit f32b4ecd26
@@ -222,12 +222,9 @@ export const Qwen36Deployment = () => {
}
}
if (hardware === 'b200') {
if (hardware === 'b200' || hardware === 'b300') {
cmd += ` \\\n --attention-backend trtllm_mha`;
}
if (hardware === 'b300') {
cmd += ` \\\n --attention-backend flashinfer`;
}
if (hwConfig.mem !== undefined) {
cmd += ` \\\n --mem-fraction-static ${hwConfig.mem}`;
}