Update README

README.md

@@ -20,9 +20,11 @@ DeepSeek-V3.2，这是一个在高计算效率与卓越推理和代理性能之
 | DTK | 25.04.1 |
 | python | 3.10.12 |
 | transformers | 4.56.1 |
+| vllm | 0.9.2+das.opt1.dtk25042 |
 | torch | 2.5.1+das.opt1.dtk25041 |
 
-推荐使用镜像:
+推荐使用镜像: image.sourcefind.cn:5000/dcu/admin/base/vllm:0.9.2-ubuntu22.04-dtk25.04.2-das1.7-py3.10-20251203
+
 - 挂载地址`-v` 根据实际模型情况修改
 ```bash
 docker run -it \
@@ -98,13 +100,14 @@ cp inference/config.json /path/to/DeepSeek-V3.2-bf16
 
 ```bash
 export VLLM_USE_V32_ENCODE=1
+
 export ALLREDUCE_STREAM_WITH_COMPUTE=1
 export VLLM_HOST_IP=x.x.x.x # 对应计算节点的IP，建议选择IB口SOCKET_IFNAME对应IP地址
 export HIP_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
 export HSA_FORCE_FINE_GRAIN_PCIE=1
 export NCCL_SOCKET_IFNAME=ibxxxx
 export GLOO_SOCKET_IFNAME=ibxxxx
-export NCCL_IB_HCA=mlx5_0:1
+export NCCL_IB_HCA=mlx5_0:1 # 根据 ibstat 查看
 unset NCCL_ALGO
 export NCCL_IB_DISABLE=0
 export NCCL_MAX_NCHANNELS=16
@@ -114,8 +117,6 @@ export NCCL_DEBUG=INFO
 export NCCL_MIN_P2P_NCHANNELS=16
 export NCCL_NCHANNELS_PER_PEER=16
 export HIP_USE_GRAPH_QUEUE_POOL=1
-export VLLM_ENABLE_MOE_FUSED_GATE=0
-export VLLM_ENFORCE_EAGER_BS_THRESHOLD=44
 export VLLM_RPC_TIMEOUT=1800000
 export VLLM_USE_FLASH_MLA=1
 
@@ -129,10 +130,6 @@ export VLLM_RANK4_NUMA=4
 export VLLM_RANK5_NUMA=5
 export VLLM_RANK6_NUMA=6
 export VLLM_RANK7_NUMA=7
-
-# BW集群需要额外设置的环境变量
-export NCCL_NET_GDR_LEVEL=7
-export NCCL_SDMA_COPY_ENABLE=0
 ```
 
 3. 启动RAY集群