Missing grid_info and grid_info explanation in docs/tutorial_deeplearning.rst

[DarioMarzella]

Describe the bug In docs/tutorial_deeplearning.rst, line 28, where generating the 'data_set', grid_info is not specified.

Actual Results or Error Info Traceback (most recent call last): File "dl_try.py", line 51, in <module> dict_filter={'IRMSD':'<4. or >10.'}) File "/home/dariomarzella/deeprank/deeprank/learn/DataSet.py", line 202, in __init__ self.process_dataset() File "/home/dariomarzella/deeprank/deeprank/learn/DataSet.py", line 277, in process_dataset self.get_grid_shape() File "/home/dariomarzella/deeprank/deeprank/learn/DataSet.py", line 813, in get_grid_shape f'Impossible to determine sparse grid shape.\n' ValueError: Impossible to determine sparse grid shape. If you are not loading a pretrained model, specify grid_shape or grid_info

Additional Context I have added the grid_info to the code block as it is in "test/test_learn.py" line 45 in commit #0ce8fd8 to branch doc_DM, but we may need a proper explanation of this feature in the tutorial.

[demainchen]

mee too! i also trap myself in this problem and no idea to resolve this problem so far

[DarioMarzella]

mee too! i also trap myself in this problem and no idea to resolve this problem so far

Hello demainchen, thanks for posting here your issue! Could I ask you exactly where did you have this problem, so that I can reproduce it exactly? While running the tutorial present in the documentation (https://deeprank.readthedocs.io/en/latest/tutorial3_learning.html ) or while running a different script you made?

In the meantime, you can check the documentation for the DataSet class. There you can find a short explanation of what grid_info and grid_shape are. I also just noticed there is a bug in how that docstring is reported, I will check it out and fix it asap.

[demainchen]

hello! Thank you for your reply so quickly, I would be appreciated if u could solve my problems. I run this code by Google Colab, and the problem occurs while I use the official tutorial document. I traped in this error when I look up the "learning part" from the document( https://deeprank.readthedocs.io/en/latest/tutorial3_learning.html#creating-training-validation-test-datasets). The error code and the total ipynb-file I will show you by attachment. Since I am not a native English speaker, sorry for my poor English. Thank u again!

Dario Marzella @.***> 于2021年6月17日周四 下午10:55写道：

mee too! i also trap myself in this problem and no idea to resolve this problem so far

Hello demainchen, thanks for posting here your issue! Could I ask you exactly where did you have this problem, so that I can reproduce it exactly? While running the tutorial present in the documentation ( https://deeprank.readthedocs.io/en/latest/tutorial3_learning.html ) or while running a different script you made?

In the meantime, you can check the documentation for the DataSet class https://deeprank.readthedocs.io/en/latest/deeprank.learn.html#module-deeprank.learn.DataSet. There you can find a short explanation of what grid_info and grid_shape are. I also just noticed there is a typo in there, I will fix it asap.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DeepRank/deeprank/issues/185#issuecomment-863310440, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASYRWNLR224F34HJOU4KGG3TTIEF5ANCNFSM4O2P4UUA .

[DarioMarzella]

Hello, it's a pleasure to help! Unfortunately, I cannot see your attachment, could you please double check you added it to your message, or re-attach it? I think the issue can be solved by simply giving the correct arguments to the grid_info, but I will be able to tell you once I see the code.

[demainchen]

[image: error_code.png] [image: error_informationn.png]thanks for your resolve my problem with patient! Can u see the picture this time?

[DarioMarzella]

Don't worry, it's a pleasure :) No, not really... can you just attach the google colab file?

In the meantime, I am trying to reproduce the bug on my side.

[demainchen]

I am very sorry for don't reply in time! this is the Link https://colab.research.google.com/drive/1HUia1liS1pd4EZtDYC3i5kAgYNgg-wFQ?usp=sharing ( https://colab.research.google.com/drive/1HUia1liS1pd4EZtDYC3i5kAgYNgg-wFQ?usp=sharing ) and the ipynb-file is attached

Dario Marzella @.***> 于2021年6月18日周五 下午7:57写道：

Don't worry, it's a pleasure :) No, not really... can you just attach the google colab file?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DeepRank/deeprank/issues/185#issuecomment-863984809, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASYRWNOAERXZ5RHYUKVMWK3TTMYD5ANCNFSM4O2P4UUA .

[DarioMarzella]

Do not worry about the reply time, we are all busy and it can always take some time to reply.

Although I am not sure why the grid_shape=[30,30,30] is not working, I think you can solve it with grid_info. Previously in your code, you defined the dictionary grid_info with number of points, resolution and atom type. You should be able to use that same dictionary to build your dataset, so instead of grid_shape=[30,30,30], use grid_info=grid_info.

EDIT: I noted my modification probably don't show up in your notebook. If you still have the grid_shape=[30,30,30] please change it to grid_info=grid_info before running it.