A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 17, 01:45 UTC
[CI] clean docker cache for neuron (#4441)
Simon Mo committed
Apr 28, 2024 at 16:32 UTC
03dd7d52bfcc4f21ba964a0cfc3fb6e7a47fb242
1 file changed
+14
.buildkite/run-neuron-test.sh
+14
@@ -4,6 +4,20 @@ set -e
4
5
# Try building the docker image
6
aws ecr get-login-password --region us-west-2 | docker login --username AWS --password-stdin 763104351884.dkr.ecr.us-west-2.amazonaws.com
7
+
8
+# prune old image and containers to save disk space, and only once a day
9
+# by using a timestamp file in tmp.
10
+if [ -f /tmp/neuron-docker-build-timestamp ]; then
11
+ last_build=$(cat /tmp/neuron-docker-build-timestamp)
12
+ current_time=$(date +%s)
13
+ if [ $((current_time - last_build)) -gt 86400 ]; then
14
+ docker system prune -f
15
+ echo $current_time > /tmp/neuron-docker-build-timestamp
16
+ fi
17
+else
18
+ echo $(date +%s) > /tmp/neuron-docker-build-timestamp
19
+fi
20
+
21
docker build -t neuron -f Dockerfile.neuron .
22
23
# Setup cleanup