Yes clearml is much better 🙂
(joking aside, mlops & orchestration in clearml is miles better)
CheerfulGorilla72 What are you looking for?
now I am still clarifying the points that have become attractive to our leadership, I will supplement my answer, How will they answer me.
You can make reports on experiments with interactive graphs
Yes, I can totally see how this is a selling point. The closest is the Project Overview (full markdown capabilities, with the ability to embed links to specific experiments). You can also add a "leader metric", so you can track the project performance/progress over time.
I have to admit that creating a better reporting tool is always pushed down in priority as I think this is a good selling point to management but the actual users would prefer other features before this one (feel free to correct me if you see it differently).
I would say that from ClearML perspective experiment management is only the beginning, the automation, data, pipeline and serving are also Tasks that need similar capabilities but are more complex, and I believe this is where ClearML shines (obviously I'm biased 🙂 )
wdyt?
is it planned to add a multicursor in the future?
CheerfulGorilla72 can you expand? what do you mean by multicursor ?
AgitatedDove14
my management liked that in wandb you can make reports on experiments with interactive graphs and that they would be viewed on the Internet + multicursor.
And I think what advantages ClearML has to compensate for this chips.
I thought that maybe someone made a comparison and find the necessary arguments in this comparison.
From the active one, we are still using local placement, logging the console, logging parameters, displaying metrics and debugging the picture with drawing pictures there.
- division into projects, subprojects, and subsubprojects with a minimal description of the key points on the subsubproject page.
AgitatedDove14 like wandb
the ability to view other values when choosing a specific one.
For example, we want to see what validation value is for a particular train loss.
yes, I agree, thank you, tell me please, is it planned to add a multicursor in the future?
for simplicity, you can make a flag for this and display only for graphs that are in sight, if it will be easier in terms of implementation or cpu costs.
it may be convenient to add an optional option to the top snippet
Oh like sync the "x-axis" between graphs?
sounds good, CheerfulGorilla72 could I ask you to open a github issue and suggest it? just so we do not forget ?
AgitatedDove14 see next:
https://github.com/allegroai/clearml/issues/630