33 lines
892 B
Bash
33 lines
892 B
Bash
export HCCL_IF_IP=141.61.39.117
|
|
export GLOO_SOCKET_IFNAME="enp48s3u1u1"
|
|
export TP_SOCKET_IFNAME="enp48s3u1u1"
|
|
export HCCL_SOCKET_IFNAME="enp48s3u1u1"
|
|
export DISAGGREGATED_PREFILL_RANK_TABLE_PATH=path-to-rank-table
|
|
|
|
export OMP_PROC_BIND=false
|
|
export OMP_NUM_THREADS=100
|
|
|
|
export VLLM_USE_V1=1
|
|
|
|
vllm serve model_path \
|
|
--host 0.0.0.0 \
|
|
--port 20002 \
|
|
--tensor-parallel-size 1\
|
|
--seed 1024 \
|
|
--served-model-name dsv3 \
|
|
--max-model-len 2000 \
|
|
---max-num-batched-tokens 2000 \
|
|
--trust-remote-code \
|
|
--gpu-memory-utilization 0.9 \
|
|
--kv-transfer-config \
|
|
'{"kv_connector": "LLMDataDistCMgrConnector",
|
|
"kv_buffer_device": "npu",
|
|
"kv_role": "kv_consumer",
|
|
"kv_parallel_size": 1,
|
|
"kv_port": "20001",
|
|
"engine_id": 0,
|
|
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_connector_v1_a3"
|
|
}' \
|
|
--additional-config \
|
|
'{"enable_graph_mode": "True"}'\
|