config cachedatset for dataset_type_performance

[wyli]

Fixes #1214

Description

update the cachedatset config to improve the performance

Checks

• Avoid including large-size files in the PR.
• Clean up long text outputs from code cells in the notebook.
• For security purposes, please check the contents and remove any sensitive info such as user names and private key.
• Ensure (1) hyperlinks and markdown anchors are working (2) use relative paths for tutorial repo files (3) put figure and graphs in the ./figure folder
• Notebook runs automatically ./runner.sh -t <path to .ipynb file>

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[Nic-Ma]

Hi @wyli ,

May I know the root cause of the performance issue? Why image_only can help improve the CacheDataset than persistentDataset?
And BTW, if you want to show runtime_cache feature in this tutorial, please update the plot figure to remove the init cache part.

Thanks.

[wyli]

Yes, the root cause is the cache init being slow, and maybe related to Project-MONAI/MONAI#5930.

image_only will remove the additional unnecessary meta_dict and the warning message. It doesn't change the relative speed comparisons.

We should keep cache_init, because the cache dataset still have the option of precompute the cache, the user can experiment and plot.

[Nic-Ma]

OK, could you please help add some description or comment in the notebook to explain the "runtime_cache" usage and init_cache plot? I feel confused when I see the plot chart which has init_cache time = 0.
Others look good to me.

Thanks in advance.

[dongyang0122]

If image_only is important to showcase advantages of PersistentDataset, I would suggest to add few words under the section of PersistentDataset explaining it? The rest of the content looks great.