miaoyibo commited on
Commit
3425837
·
1 Parent(s): eca802f
Files changed (1) hide show
  1. serve_vllm.sh +2 -2
serve_vllm.sh CHANGED
@@ -3,8 +3,8 @@
3
  python -m vllm.entrypoints.openai.api_server \
4
  --model moonshotai/Kimi-Dev-72B \
5
  --tensor-parallel-size 4 \
6
- --max-num-seqs 8 \
7
- --max-model-len 131072 \
8
  --gpu-memory-utilization 0.9 \
9
  --host localhost \
10
  --served-model-name kimi-dev \
 
3
  python -m vllm.entrypoints.openai.api_server \
4
  --model moonshotai/Kimi-Dev-72B \
5
  --tensor-parallel-size 4 \
6
+ --max-num-seqs 4 \
7
+ --max-model-len 65536 \
8
  --gpu-memory-utilization 0.9 \
9
  --host localhost \
10
  --served-model-name kimi-dev \