EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Doc] Modify DeepSeek-R1/V3.1 documentation (#5426)

Browse Source 
### What this PR does / why we need it?
Modify DeepSeek-R1/V3.1 documentation. Mainly update the mtp size and some other configs.

Signed-off-by: GDzhu01 <809721801@qq.com>
This commit is contained in:
Zhu Yi Lin
2025-12-27 17:13:58 +08:00
committed by GitHub
parent 09f71c14a6
commit 04104031d0
2 changed files with 15 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
docs/source/tutorials/DeepSeek-R1.md
View File

@@ -88,12 +88,14 @@ local_ip="xxxx"
# AIV
export HCCL_OP_EXPANSION_MODE="AIV"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export VLLM_ASCEND_ENABLE_MLAPO=1
export VLLM_ASCEND_BALANCE_SCHEDULING=1
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_USE_MODELSCOPE=True
vllm serve vllm-ascend/DeepSeek-R1-W8A8 \
@@ -143,6 +145,9 @@ export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=1
export HCCL_BUFFSIZE=200
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
export VLLM_ASCEND_BALANCE_SCHEDULING=1
@@ -167,8 +172,8 @@ vllm serve vllm-ascend/DeepSeek-R1-W8A8 \
--max-model-len 16384 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.94 \
--speculative-config '{"num_speculative_tokens":1,"method":"mtp"}' \
--gpu-memory-utilization 0.92 \
--speculative-config '{"num_speculative_tokens":3,"method":"mtp"}' \
--compilation-config '{"cudagraph_capture_sizes":[4,16,32,48,64], "cudagraph_mode": "FULL_DECODE_ONLY"}'
```
@@ -187,6 +192,9 @@ export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=1
export HCCL_BUFFSIZE=200
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
export VLLM_ASCEND_BALANCE_SCHEDULING=1
@@ -213,8 +221,8 @@ vllm serve vllm-ascend/DeepSeek-R1-W8A8 \
--max-model-len 16384 \
--max-num-batched-tokens 4096 \
--trust-remote-code \
--gpu-memory-utilization 0.94 \
--speculative-config '{"num_speculative_tokens":1,"method":"mtp"}' \
--gpu-memory-utilization 0.92 \
--speculative-config '{"num_speculative_tokens":3,"method":"mtp"}' \
--compilation-config '{"cudagraph_capture_sizes":[4,16,32,48,64], "cudagraph_mode": "FULL_DECODE_ONLY"}'
```

7
docs/source/tutorials/DeepSeek-V3.1.md
View File

@@ -166,8 +166,7 @@ export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export VLLM_USE_V1=1
export OMP_NUM_THREADS=1
export HCCL_BUFFSIZE=200
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
@@ -220,7 +219,7 @@ export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export OMP_NUM_THREADS=1
export HCCL_BUFFSIZE=200
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export VLLM_ASCEND_ENABLE_MLAPO=1
@@ -249,7 +248,7 @@ vllm serve /weights/DeepSeek-V3.1-w8a8-mtp-QuaRot \
--trust-remote-code \
--no-enable-prefix-caching \
--gpu-memory-utilization 0.92 \
--speculative-config '{"num_speculative_tokens": 1, "method": "mtp"}' \
--speculative-config '{"num_speculative_tokens": 3, "method": "mtp"}' \
--compilation-config '{"cudagraph_capture_sizes":[4,16,32,48,64], "cudagraph_mode": "FULL_DECODE_ONLY"}'
```