-
For allowing humans to better distinguish between jobs (and search through them), add an optional `description` text field to the `job` model.
-
Job executors need to mark when they started a job and when they finished one. This could be achieved by simply setting the `start_time` and `end_time` fields in the job table. But this does not cover…
-
If a job fails due to a job template containing an error, if would be useful being able to re-run the failed job again (after the issue has been fixed). Re-running the job would remove error, start/en…
-
* adding a job template to the backend stores `options` for `input` and `default` for `parameter`
* when loading a job template from the backend, `options` and `parameter` are empty
* the serializer…
-
* add ability to easily add migrations for job templates (per domain)
* job template migrations are stored in json format to make editing easy, resembling create request for job templates
* any fiel…
-
* not all docker images will work with all hardware generations (of graphics cards): not all PyTorch versions work with 1080 Ti versions, despite CUDA being compatible
* the `ufdl-core-app/src/ufdl/c…
-
setup.py dependency on pycocotools from Github fails with:
```
ERROR: Packages installed from PyPI cannot depend on packages which
are not also hosted on PyPI.
wai.annotations depends on pycocot…
-
* create model and API end point for querying/updating the CUDA - NVIDIA driver version relation
* the `ufdl-core-app/src/ufdl/core_app/migrations/cuda` directory contains CSV file for migration and …
-
More filters:
* gt - greater than
* ge - greater or equal
* lt - less than
* le - less or equal
-
in order to list jobs that haven't been started yet (`start_time` is null), an additional filter for listing rows with nulls in a field is necessary, e.g., called `is_null`