Eren Balatkan

Bump, forced label results in misplaced element when we have no label. For example, In my case I'm displaying list of elements where you can modify value of each element....

I would like to thank you for sharing this aswell. In machine learning it is generally really difficult to estimate how much data you need to get a good model,...

> i managed to convert the model to torch script with trace, and serve it with triton, but still struggle to convert it to onnx. Greetings, Im having trouble converting...

I also would like to make a suggestion in this regard, ClearML does not seem to automatically pick up new files if they are not added to Git (even if...

Wouldn't it be possible to treat project folders similar to the way datasets work? I'm not sure about how git diff part could be handled.

Yeah, Git support is already pretty good from what I have tested so far, but ability to run projects that are not connected to Git would be a huge plus...

> I think this is doable, would zipping the entire (really the entire, no filtering added) working directory work ? > How would one specify the "root folder" ? (e.g....

Looks perfect for me. This would give us a ton more flexibility with experiments.

From my perspective, stored code must be logged to ClearML task to ensure that every experiment is reproducable, currently saving Git commit already ensures this but for this alternative approach...

Hello, I would suggest using ```sudo clearml-data close --verbose``` Verbose output may give you a better hint as into why it is taking long.