Running prodigy with docker

We are running into some problems running prodigy with docker. The docker image uses centos6 with Python 3.6. As given in the user guide, we have changed “host” to “0.0.0.0” in prodigy.json. The below commands work fine when they are run locally, but they fail in docker.

  1. This command occasionally starts a web server but most of the time it results in a “Killed” state.
    CMD:
    prodigy ner.manual test_dataset en_core_web_lg prodigy-module/queries.txt --label TEST
    RESULT:
    Using 1 labels: TEST
    /opt/rh/rh-python36/root/usr/bin/prodigy: line 1: 25 Killed python -m prodigy “$@”

  2. This command always fails.
    CMD:
    prodigy terms.teach test_dataset en_vectors_web_lg --seeds “dog,cat,ant”
    RESULT:
    /opt/rh/rh-python36/root/usr/bin/prodigy: line 1: 31 Segmentation fault python -m prodigy “$@”

What’s the memory limit on the container? It sounds like you’re running into resource constraints.

1 Like

It was set to 1 GB. It is working now after increasing the memory.
Thank you! :slight_smile:

1 Like