Open denisvmedyantsev opened 1 month ago
Backport to 0.4
Changed compose file to faster testing:
media_file.sh
to video_loop.sh
(to avoid restarting containers);SYNC_OUTPUT=False
and EOS_ON_LOOP_END=True
.services:
video-source-camera-1:
# image: ghcr.io/insight-platform/savant-adapters-gstreamer:0.4.7
image: ghcr.io/insight-platform/savant-adapters-gstreamer:latest
# restart: unless-stopped
volumes:
- zmq_sockets:/tmp/zmq-sockets
- ./downloads_cache:/tmp/video-loop-source-downloads
- ./data:/data
environment:
- LOCATION=/data/cam1.mp4
- SOURCE_ID=camera-1
- DOWNLOAD_PATH=/tmp/video-loop-source-downloads
- ZMQ_ENDPOINT=dealer+connect:ipc:///tmp/zmq-sockets/input-video.ipc
- FILE_TYPE=video
- READ_METADATA=FALSE
- SYNC_OUTPUT=False
- EOS_ON_LOOP_END=True
entrypoint: /opt/savant/adapters/gst/sources/video_loop.sh
depends_on:
module:
condition: service_healthy
module:
# image: ghcr.io/insight-platform/savant-deepstream:0.4.7-6.4
image: ghcr.io/insight-platform/savant-deepstream:latest
volumes:
- zmq_sockets:/tmp/zmq-sockets
entrypoint: >
bash -c "echo '
name: test-module
parameters:
batch_size: 1' | python -m savant.entrypoint
"
environment:
- LOGLEVEL=info
- ZMQ_SRC_ENDPOINT=router+bind:ipc:///tmp/zmq-sockets/input-video.ipc
- ZMQ_SINK_ENDPOINT=pub+bind:ipc:///tmp/zmq-sockets/output-video.ipc
- METRICS_TIME_PERIOD=30
runtime: nvidia
volumes:
zmq_sockets:
On branch bugfixes/04x looks like the time between processing EOS and reconnection of the source is increasing. On develop there's another issue: the module stops processing frames after some time (1-2 minutes).
I used 4 videos (30+30+25+25=110 FPS) and a dump module
FPS logs on start
after ~1 hour
docker-compose.yml