BTW, I suggest for new questions, just ask in the clearml-community. I'm really happy to help but I almost missed this message 😄

OutrageousSheep60 The python package is in testing. Hopefully will be out Sunday \ Monday :)

VexedCat68 you mean the artifact in the previous step is called "merged_dataset_id"? Is it an artifact or is it a parameter? And what issues are you having with accessing the parameter?

And as for clearml-data I would love to have more examples but not 100% sure what to focus on as using clearml-data is a bit...simple? In my, completely biased, eyes. I assume you're looking for workflow examples, and would love to get some inspiration 🙂

And in the pre_execute_callback, I can access this:
a_pipeline._nodes[a_node.parents[0]].job.task.artifacts['data_frame']

pipe._nodes['stage_data'].job.task.artifacts

The new welcome screen to pipelines and our fancy new icon on the left sidebar 😄

And some real pipeline (As real as our tests get 😄 )

EnviousStarfish54 BTW, as for absolute reproducibility, you are obviously right. If you use S3 to store the data, and you changed the data in S3 then we can't catch it.
Our design compresses (zips) the files and store them in a version somewhere. If this is modified than you are trying hard to break stuff 🙂 (Although you can). This is not the most efficient space-wise when it comes to images \ videos, for these, you can save links, but I think it's only in the enterprise version but then,...

Hi EnviousStarfish54 If you want to not send info to the server, I suggest you to set an environment variable, this way as long as the machine has this envvar set it won't send to the server

EnviousStarfish54 VivaciousPenguin66 Another question if we're in a sharing mood 😉 Do you think a video \ audio session with one of our experts, where you present a problem you're having (let's say large size of artifacts) and he tries to help you, or even can give some example code \ code skeleton. Would something like that be of interest? Would you spend some time in such monthly session?

As we always say, you came because it's free, you stayed because features are being released before git issues are even opened 😉
Thanks for contributing back with ideas and inputs! 😄

EnviousStarfish54 VivaciousPenguin66 So for random seed we have a way to save it so this should be possible and reproducible.
As for execution progress I totally agree. We do have our pipelining solution but I see it's very common to use us only for experiment tracking and use other tools for pipelining as well.
Not trying to convert anyone but may I ask why did you choose to use another tool and not the built-in pipelining feature in ClearML? Anything missing? Or did you just build the in...

CLEARML_NO_DEFAULT_SERVER=1

LOL Love this Thread and sorry I didn't answer earlier!

VivaciousPenguin66 EnviousStarfish54 I totally agree with you. We do have answers to "how do you do X or Y" but we don't have workflows really.

What would be a logical place to start? Would something like "training a Yolo V3 person detector on COCO dataset and how you enable continuous training (let's say adding PASCAL dataset afterwords) be something interesting?

The only problem is the friction between atomic and big picture. In...

DilapidatedDucks58 , We have a hunch we know what's wrong (we think we treat loading data like loading model and then we register each file \ files pickle as a model which takes time). How are you loading data? Is monai built inside pytorch? Or are you downloading it and loading manually? If you can share the loading code that might be helpful 🙂

Hi Binoy, At the moment, we only support this featured in our enterprise offering. We're now adding more volume to our paid tier and this is indeed a candidate feature to be added. Stay tuned 🙂

Yeah, it might be the cause...I had a script with OOM and it crashed regularly 🙂

BTW! can you elaborate on the need for elevated privileges? What can't he do that you want him to?

got it 🙂

Sure 😄

That's right. once you call clearml-data close, the completed dataset is immutable. This is a very important feature if traceability is important as once an experiment uses a dataset version, we want to make sure it doesn't change without leaving a trace!

We'll check this. I assume we don't catch the error somehow or the proccess doesn't indicate it died failing

I think that's a hydra issue 🙂 I was able to reproduce this locally. I'll see what can be done

I think the best model name is person_detector_lr0.001_batchsz32_accuracy0.63.pkl 😄

pipe.add_step(name='stage_process', parents=['stage_data', ],
base_task_project='examples', base_task_name='pipeline step 2 process dataset',
parameter_override={'General/dataset_url': '${stage_data.artifacts.dataset.url}',
'General/test_size': 0.25}, pre_execute_callback=pre_execute_callback_example, post_execute_callback=post_execute_callback_example)

Hi Moki, Great idea! We'll add it to our plans and update here once it's done 😄

Hi JumpyPig73 , I reproduced the OOM issue but for me it's failing. Are you handling the error in python somehow so the script exists gracefully? otherwise it looks like a regular python exception...

🙂

Happy our intention was still clear