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

[Doc][P/D] Fix MooncakeConnector's name (#5172)

Browse Source 
### What this PR does / why we need it?
vLLM community has integrated their MooncakeConnector. The original
scripts will now find this MooncakeConnector instead of the one from
vLLM-Ascend. All scripts that involve using the MooncakeConnector need
to be modified to another name.

### Does this PR introduce _any_ user-facing change?
Yes, users need to use a new name to load vLLM-Ascend MooncakeConnector.

### How was this patch tested?
By CI.

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: nwpu-zxr <zhouxuerong2@huawei.com>
This commit is contained in:
zxr2333
2025-12-18 22:29:19 +08:00
committed by GitHub
parent 2304218f90
commit 073a3a6e6c
12 changed files with 35 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
docs/source/tutorials/pd_disaggregation_mooncake_multi_node.md
View File

@@ -504,7 +504,7 @@ vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
--additional-config '{"recompute_scheduler_enable":true,"enable_shared_expert_dp": true}' \
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
'{"kv_connector": "MooncakeConnectorV1",
"kv_role": "kv_producer",
"kv_port": "30000",
"engine_id": "0",
@@ -564,7 +564,7 @@ vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
--additional-config '{"recompute_scheduler_enable":true,"enable_shared_expert_dp": true}' \
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
'{"kv_connector": "MooncakeConnectorV1",
"kv_role": "kv_producer",
"kv_port": "30100",
"engine_id": "1",
@@ -625,7 +625,7 @@ vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
--additional-config '{"recompute_scheduler_enable":true,"multistream_overlap_shared_expert": true,"lm_head_tensor_parallel_size":16}' \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
'{"kv_connector": "MooncakeConnectorV1",
"kv_role": "kv_consumer",
"kv_port": "30200",
"engine_id": "2",
@@ -685,7 +685,7 @@ vllm serve /path_to_weight/DeepSeek-r1_w8a8_mtp \
--additional-config '{"recompute_scheduler_enable":true,"multistream_overlap_shared_expert": true,"lm_head_tensor_parallel_size":16}' \
--compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}' \
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
'{"kv_connector": "MooncakeConnectorV1",
"kv_role": "kv_consumer",
"kv_port": "30200",
"engine_id": "2",