DepressedChimpanzee34 ClearML tries to conserve storage by limiting the history length for debug images (see sdk.metrics.file_history_size
https://clear.ml/docs/latest/docs/configs/clearml_conf#sdk-section ), though the history can indeed grow large by setting a large value or using a metric/variant naming scheme to circumvent this limit.
Does your use case call for accessing a specific iteration for all images or when looking at a specific image? Note that the debug image viewer (wh...
UnsightlySeagull42 The upgrade process is slightly different depending on the environment in which you've deployed your ClearML server (e.g. for a https://allegro.ai/clearml/docs/docs/deploying_clearml/clearml_server_linux_mac.html#upgrading ).
Note the document you are referring to only applies once when you're moving from the older pre-0.16 versions in which case DB migration is required.
If your server is more up to date (0.16 and newer) you should be OK with the link above.
WittyOwl57 I just used a couple of the experiments in the https://app.community.clear.ml/projects/764d8edf41474d77ad671db74583528d/ of the free tier server.
ItchyJellyfish73 Have you looked at the https://clear.ml/docs/latest/docs/clearml_agent#dynamic-gpu-allocation ?
WittyOwl57 The UI shows repo and package detailed comparison under the "Details"/"Execution" (See sample screenshot), whereas auto-logged environment variables are shown under the "HyperParameters" comparison tab.
What do you find missing beyond those?
TightElk12 This makes a lot of sense - should make it into one of the coming releases
@<1523706095791509504:profile|FiercePenguin76> The "Log" tab has been renamed "Console" in ClearML 0.17.0 - Thanks for pointing out the outdated description.
ScrawnyLion96 Looks like a case of broken links - Check out https://clear.ml/docs/latest/docs/references/api/definitions#tasksexecution and https://clear.ml/docs/latest/docs/references/api/definitions#tasksconfiguration_item
CooperativeSealion8 For future reference, notice there's a configuration reference available at https://allegro.ai/docs/references/trains_ref/
Thanks for noticing @<1523708920831414272:profile|SuperficialDolphin93> - ClearML is already there under it's legacy "Trains" name, it's indeed past time for an update.
WittyOwl57 No worries 🙂 happens to the best!
OutrageousSheep60 You can see https://github.com/allegroai/clearml/issues/724 a discussion on the topic.
TL;DR:
Currently the containing project is available in the UI as a tooltip to the dataset name An alternate "Project view" to the datasets page is in the works
@<1687643893996195840:profile|RoundCat60> Looks like the docs have not caught up yet with recent structure change in the repo which renamed the 'server' folder to 'apiserver'.
So... the correct link would be None
BattyLion34 Adding to AgitatedDove14 hint. See the following docs page: https://allegro.ai/clearml/docs/docs/deploying_clearml/clearml_config_for_clearml_server.html
@<1523709410411548672:profile|NuttyFox2> Since the default server user configuration does not require authentication, I'm assuming your use case calls for some users being authenticated where others are not?
Such mixed access mode is currently not on the near term roadmap for the OSS server - You should create a feature request to help push it into the development plan.
DepressedChimpanzee34 Always appreciated
KindGiraffe71 Have you checked out the https://github.com/allegroai/clearml/blob/master/examples/frameworks/pytorch-lightning/pytorch_lightning_example.py ? https://clearml.slack.com/archives/CTK20V944/p1616070536033700 previous discussion provides some insight into how it works under the hood.
DepressedChimpanzee34 Apologies for missing your previous comment.
Totally agree that the global selection indicator should maintain its 'clear selection' behaviour even if some/all of the selection is off-screen.
IrateDolphin19 ClearML provides for saving files generated as part of your code execution through the https://clear.ml/docs/latest/docs/references/sdk/task#upload_artifact . For your use case, you can have your code thus create the artifact as it runs, you can set the specific storage location when you edit your configuration, through the task's output_uri field.
Does this help?
RotundHedgehog76 Have you tried clearml-data add --files .
? (Probably best to try on a smaller subset first)
GreasyPenguin14 When the project description is empty you get a "Add project overview" instead if the "Edit" button:
DepressedChimpanzee34a filter similar to one in the scalars page where you can display a subset of the reported debug images can be useful
The scalars page provides a metric hide/show control - Is this the one you mean? The debug images page also provides a filter by metric - Depending on your naming policy this can easily be used to focus on more sparsely appearing images.
Else, an example of the filter you were thinking of would be appreciated.
Regardless, direct iteration access cou...
@<1523701157564780544:profile|TenseOstrich47> The storage in question here is what's available on the machine hosting the ClearML server's docker containers (specifically, the ES one).
@<1523701157564780544:profile|TenseOstrich47> Seems like the ClearML website is temporarily down 😞 . Should be resolved soon though.
DefeatedCrab47 Happy you're finding Trains useful 🙂
but it definitely has it's advantages if TRAINS would support it (early stage Data Science infrastructure).
No doubt, and I definitely see such usable example in the cards for Trains' upcoming versions...
@<1523705301990117376:profile|WickedCat12> ClearML Scalars explicitly show metrics time progression (you can display iteration/wall-time).
Plotting one metric against another is a feature that lies further down ClearML's roadmap.
If your metric is reported only once per epoch you can make use of the existing scalars functionality by making use of the iteration parameter when reporting your metric to reflect the epoch instead.
Does this make sense?
DepressedChimpanzee34 Thanks for clarifying where the current debug images display falls short for your use case - Extending the filtering to liken the behaviour of the scalars sound like a great idea 🙂
From the https://github.com/allegroai/trains-server/releases/tag/0.13.0 :
Reports average load metrics per day (CPU/memory) Reports average workload per day (amount and average duration of queues, agents and experiments)