Until today everything works well. I have change the model to antelopev2 and redo facial recognition. Everything has worked fine. Then I uploaded a new set (about 3500 assets) and launch the process of "face detection" and "facial recognition" after the process of creating thumbnails images and metadats extraction. Then I got the following error at every asset treatment :
immich_server | [Nest] 7 - 03/06/2025, 10:08:14 PM WARN [Microservices:MachineLearningRepository] Machine learning request to "http://immich-machine-learning:3003" failed: fetch failed
immich_server | [Nest] 7 - 03/06/2025, 10:08:14 PM ERROR [Microservices:{"id":"40dcafbd-d371-46d7-8d31-9279b08f0b2f"}] Unable to run job handler (faceDetection/face-detection): Error: Machine learning request '{"facial-recognition":{"detection":{"modelName":"antelopev2","options":{"minScore":0.7}},"recognition":{"modelName":"antelopev2"}}}' failed for all URLs
In "ps ax" I have this for the ML process command :
/opt/venv/bin/python /opt/venv/bin/gunicorn app.main:app -k app.config.CustomUvicornWorker -c gunicorn_conf.py -b [::]:3003 -w 1 -t 120 --log-config-json log_conf.json --keep-alive 2 --graceful-timeout 0
In conf I have this as parameter : http://immich-machine-learning:3003
I notice that the parameter value is not in the command line of the process. Bug ?
Thanks for the help.
.