engine = "sglang"
name = "Minimax-M3-v0 NVFP4"
docker_image = "scitrera/dgx-spark-sglang-mm:v0"
model = "sparkarena/Minimax-M3-v0-NVFP4"
port = 30000
required_nodes = 4
tensor_parallel_size = 4
distributed_backend = "mp"

entrypoint = ["sglang", "serve"]

args = [
    "--mem-fraction-static", "0.8",
    "--context-length", "200000",
    "--chunked-prefill-size", "8192",
    "--fp4-gemm-backend", "cutlass",
    "--load-format", "instanttensor",
    "--kv-cache-dtype", "auto",
    "--reasoning-parser", "minimax-m3",
    "--tool-call-parser", "minimax-m3",
    "--moe-runner-backend", "triton",
    "--trust-remote-code",
    "--cuda-graph-max-bs", "16",
    "--cuda-graph-bs", "1", "2", "4", "8", "12", "16",
    "--host", "0.0.0.0",
]

[env]
TORCH_NCCL_ASYNC_ERROR_HANDLING = "1"