Maisi VAE notebook

[Can-Zhao]

Fixes # .

Description

Add MAISI notebook and adjust it to a setting that can run on 16G V100.

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

[mingxin-zheng]

Hi @Can-Zhao , the CI indicates the something may be wrong in the last two cells. Can you try running it locally?

Update: I finish running the notebook locally. The last cell fails probably due to out-of-memory in github CI V100 16G. Last mem use I capture during I run the notebook
2024/07/12 15:43:44.493, 16640 MiB

[mingxin-zheng]

I think the PR overall looks good to me. It would be better to keep the outputs of the last 3 cells of the notebook.

[Can-Zhao]

I think the PR overall looks good to me. It would be better to keep the outputs of the last 3 cells of the notebook.

May I ask if I need to finish the whole training process to generate the output cell?

[KumoLiu]

I think the PR overall looks good to me. It would be better to keep the outputs of the last 3 cells of the notebook.

May I ask if I need to finish the whole training process to generate the output cell?

Yes, and after that you can reset the epoch to a small number such as 2 then the ci will not run the whole training.

[Can-Zhao]

I think the PR overall looks good to me. It would be better to keep the outputs of the last 3 cells of the notebook.

May I ask if I need to finish the whole training process to generate the output cell?

Yes, and after that you can reset the epoch to a small number such as 2 then the ci will not run the whole training.

That takes a lot of time, I do not have another week...I did not use these datasets for released model anyway...

[Can-Zhao]

Hi @Can-Zhao , the CI indicates the something may be wrong in the last two cells. Can you try running it locally?

Update: I finish running the notebook locally. The last cell fails probably due to out-of-memory in github CI V100 16G. Last mem use I capture during I run the notebook 2024/07/12 15:43:44.493, 16640 MiB

I changed configs and it is resolved

[Can-Zhao]

I think the PR overall looks good to me. It would be better to keep the outputs of the last 3 cells of the notebook.

Added output cells for 2 epochs