Seppo Enarvi
Seppo Enarvi
By "General PL module for training any custom SSL model" do you need a command line application?
Currently the SSL models take the name of the backbone, such as "resnet18", in the constructor. I would suggest that they also accept any network (`nn.Module`) that will be used...
@Atharva-Phatak I have some questions. 1) Is the ``CustomLightningModule`` class something that the user would write, and Bolts would offer only the heads and criterions? Or do you suggest that...
I didn't know about [docker-compose.override.yml](https://docs.docker.com/compose/extends/#multiple-compose-files). That can also be used to modify the configuration without having to modify `docker-compose.yml` (which will be overwritten by the menu).
I managed to set the timezone with TZ to every service that I use, where it makes a visible change to the user, i.e. in the UI. Is the wrong...
Right, at least it affects the timestamps in the logs. Great work!
> I didn't need to roll my own Dockerfile... passing this as the command works (I'm starting up the mlflow server in a docker compose): > > `bash -c "python3...
Here's a PR with some documentation changes: https://github.com/Lightning-AI/litdata/pull/264 I didn't quite understand how the placeholder would work.
@tchaton something like that could be an elegant solution. I guess then you need to write a serializer for the `Placeholder` class. Was your idea that all the elements of...
I don't have the possibility at the moment. I was able to easily work around the problem by using NumPy arrays instead of Python lists. Now I'll have to move...