when starting the test training i get unknown error
KiriUbu opened this issue · 4 comments
When i start the test training i get the Error "Unknown Error"
I had no errors during the installation. I am using Ubuntu 18.04.4 LTS what a i doing wrong?
thank you for every suggestion
A good way to debug the problem would be to check the api container logs.
Open your terminal and type docker ps you should see a docker container named docker belonging to an image called tf_docker_sdk
Take the id of that container and type docker logs the id you just copied.
This should give you a clearer idea on why the training failed.
Thank you i will try it.
Does somebody have solve it?
Error in my sdk docker logs is absent until I click "next" button second time (in this case error tell me that container name is already taken), so I guess it should appear somewhere else.
Only uncommon action I performed is running containers via terminals instead of run.yml file as far as runtime string is not supported by newest version of docker-compose
There is commands which I run from 3 different terminals (~/TrainingAPI/ is my API root directory)
sudo docker run --rm -it --gpus all --net=host -v /var/run/docker.sock:/var/run/docker.sock -v ~/TrainingAPI/datasets:/datasets -v ~/TrainingAPI/training_api/networks_weights:/weights -v ~/TrainingAPI/checkpoints:/checkpoints tf_docker_sdk
sudo docker run --rm -it -v /var/run/docker.sock:/var/n/docker.sock -v ~/TrainingAPI/inference/models:/models -v ~/TrainingAPI/inference/src/main:/main -p 4343:4343 tf_inference_api_for_training_gui
sudo docker run --rm -it -p 4200:80 tf_training_ui
Is there something else I should provide?
Also must notice, that new job appears in the job list, and I even could open logs window.
UPD: There is my logs window before error
I see that you are using firefox as your browser. We recommend to try a different browser (e.g. chrome, chromium or opera) and see if the problem persists