agg_router.sh 1.24 KB
Newer Older
1
2
3
4
5
6
7
#!/bin/bash
# SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

# Setup cleanup trap
cleanup() {
    echo "Cleaning up background processes..."
8
9
    kill $DYNAMO_PID $WORKER_PID 2>/dev/null || true
    wait $DYNAMO_PID $WORKER_PID 2>/dev/null || true
10
11
12
13
14
    echo "Cleanup complete."
}
trap cleanup EXIT INT TERM

# run clear_namespace
15
python3 -m dynamo.sglang.utils.clear_namespace --namespace dynamo
16
17

# run ingress
18
python -m dynamo.frontend --router-mode kv --http-port=8000 &
19
20
21
DYNAMO_PID=$!

# run worker
22
python3 -m dynamo.sglang \
23
24
25
26
27
  --model-path deepseek-ai/DeepSeek-R1-Distill-Llama-8B \
  --served-model-name deepseek-ai/DeepSeek-R1-Distill-Llama-8B \
  --page-size 16 \
  --tp 1 \
  --trust-remote-code \
28
  --skip-tokenizer-init \
29
30
31
  --kv-events-config '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:5557"}' &
WORKER_PID=$!

32
CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.sglang \
33
34
35
36
37
38
39
  --model-path deepseek-ai/DeepSeek-R1-Distill-Llama-8B \
  --served-model-name deepseek-ai/DeepSeek-R1-Distill-Llama-8B \
  --page-size 16 \
  --tp 1 \
  --trust-remote-code \
  --skip-tokenizer-init \
  --kv-events-config '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:5558"}'