Martin.B

@swigicat > Could you please enable text selection or a true html / TeX table copy/paste button? The JSON export is not very useful right now for this job. I...

Hi @idantene , thanks for the bump! Recapping here: 1. Report a single value (not time series), might be a single "key" (as opposed to the scalar pair title+series) 2....

> Maybe an option to choose if a reported value is optimized to be maximal or minimal first (i.e. highlight larger F1 scores, but lower MSE scores...) i.e. sort the...

Hi @vlad-ivanov-name > How important is os.fork patching for experiment tracking? In short very 🙂 The issue is that without addressing the fork issue, all the "original" threads and thread...

Quick update, v1.4.1 is out with a fix 😄 `pip install clearml==1.4.1` @vlad-ivanov-name > So would it be correct to say this patch exists to capture stdout/stderr of forked processes?...

> Another suggestion would be to have a config option to ignore Git completely and just store entire project folder for experiment. @markus289 and @ErenBalatkan unfortunately I'm not actually sure...

> Wouldn't it be possible to treat project folders similar to the way datasets work? Yes it would, but it also means we need to add support for the `clearml-agent`...

> ... but ability to run projects that are not connected to Git would be a huge plus for us. I think this is doable, would zipping the entire (really...

Following the request in #395 for a more specific control over the data that will be packaged. We could try something like ```python task.package_repo(root_folder='.', include_files=['*.py', '*.txt'], exclude_files=['*.pyc']) ``` Also introduce...

> ... as if they're uploaded using task.upload_artifact? Basically the way I imagine it is, we zip everything (source code , data, etc.), then put it as an artifact on...