INNER CODE UNIT · Python
max_request_bytes
raullenchai/Rapid-MLX · vllm_mlx/models/deepseek_v41_native/server.py:37
max_request_bytes: int = 8 * 1024 * 1024,
body_receive_timeout_seconds: float = 15.0,
default_timeout: float = 1800.0,
max_concurrent_requests: int = 256,
cors_policy: Any | None = None,
reasoning_parser_name: str | None = "deepseek_v4",
) -> None:
import uvicorn
from vllm_mlx.speculative.dflash.server import _build_app, _dflash_executor
require_product_memory()
target_path = download_target_snapshot()
mtp_path = download_mtp_snapshot()
def _load_all():
return load_product_runtime(
str(target_path),