[Frontend][Bugfix] support prefill decode disaggregation on deepseek (#14824)

Signed-off-by: billishyahao <bill.he@amd.com>
Co-authored-by: Zhai Feiyue <80079571+ZhaiFeiyue@users.noreply.github.com>
This commit is contained in:
billishyahao
2025-03-20 00:00:33 -07:00
committed by GitHub
co-authored by Zhai Feiyue
parent bfe2fe0af4
commit 742369d35a
3 changed files with 70 additions and 21 deletions
@@ -8,6 +8,9 @@ set -xe
echo "🚧🚧 Warning: The usage of disaggregated prefill is experimental and subject to change 🚧🚧"
sleep 1
# meta-llama/Meta-Llama-3.1-8B-Instruct or deepseek-ai/DeepSeek-V2-Lite
MODEL_NAME=${HF_MODEL_NAME:-meta-llama/Meta-Llama-3.1-8B-Instruct}
# Trap the SIGINT signal (triggered by Ctrl+C)
trap 'cleanup' INT
@@ -44,18 +47,20 @@ wait_for_server() {
# You can also adjust --kv-ip and --kv-port for distributed inference.
# prefilling instance, which is the KV producer
CUDA_VISIBLE_DEVICES=0 vllm serve meta-llama/Meta-Llama-3.1-8B-Instruct \
CUDA_VISIBLE_DEVICES=0 vllm serve $MODEL_NAME \
--port 8100 \
--max-model-len 100 \
--gpu-memory-utilization 0.8 \
--trust-remote-code \
--kv-transfer-config \
'{"kv_connector":"PyNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2}' &
# decoding instance, which is the KV consumer
CUDA_VISIBLE_DEVICES=1 vllm serve meta-llama/Meta-Llama-3.1-8B-Instruct \
CUDA_VISIBLE_DEVICES=1 vllm serve $MODEL_NAME \
--port 8200 \
--max-model-len 100 \
--gpu-memory-utilization 0.8 \
--trust-remote-code \
--kv-transfer-config \
'{"kv_connector":"PyNcclConnector","kv_role":"kv_consumer","kv_rank":1,"kv_parallel_size":2}' &
@@ -78,7 +83,7 @@ sleep 1
output1=$(curl -X POST -s http://localhost:8000/v1/completions \
-H "Content-Type: application/json" \
-d '{
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
"model": "'"$MODEL_NAME"'",
"prompt": "San Francisco is a",
"max_tokens": 10,
"temperature": 0
@@ -87,7 +92,7 @@ output1=$(curl -X POST -s http://localhost:8000/v1/completions \
output2=$(curl -X POST -s http://localhost:8000/v1/completions \
-H "Content-Type: application/json" \
-d '{
"model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
"model": "'"$MODEL_NAME"'",
"prompt": "Santa Clara is a",
"max_tokens": 10,
"temperature": 0