[MLX] bench_one_batch: thread --quantization through to MlxModelRunner (#25221)

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-14 09:56:42 -07:00
parent 90afd680f3
commit 3fc60e5748
1 changed files with 1 additions and 0 deletions
@@ -531,6 +531,7 @@ class _MlxBenchRunner:
            trust_remote_code=server_args.trust_remote_code,
            disable_radix_cache=True,
            mem_fraction_static=server_args.mem_fraction_static,
+            quantization=server_args.quantization,
        )
        if server_args.max_total_tokens is not None:
            init_kwargs["pool_size"] = server_args.max_total_tokens