How to fix "the kernel appears to have died. It will restart automatically."?

Help
1

Hi all,

I follow a tutorial that run on Jupyter Notebook:

   scvi.model.SCVI.setup_anndata(adata)
    vae = scvi.model.SCVI(adata)
    vae.train()

“The kernel appears to have died. It will restart automatically.”

GPU available: False, used: False
TPU available: False, using: 0 TPU cores
IPU available: False, using: 0 IPUs
HPU available: False, using: 0 HPUs
Epoch 1/400: 0%| | 0/400 [00:00<?, ?it/s]

This error seems because pytorch. I run 2 code samples to test pytorch, one is OK, another require upgrading MacOS to 12.3. What is wrong here? Thank you so much.

2

Do you have enough memory? Does pytorch work at all outside the context of this notebook?

1 Like
3

One sample work and another require upgrading MacOS to 12.3. You mean the Ram or hard drive storage?

4

My jupyter lab also says kernel died after running(python 3.9.12, centos7, 128gb ram, 500gb hdd):

sc.settings.verbosity = 3             # verbosity: errors (0), warnings (1), info (2), hints (3)
sc.logging.print_header()
sc.settings.set_figure_params(dpi=120, facecolor='white')

After comment out

#sc.logging.print_header()

notebook works but nothing printed out.

5

I run on GPU and don’t get that error anymore.

6

My server doesn’t have a GPU. GPU is required?

7

I am not sure but GPU worked in the cases I know.

8

As @adamgayoso hinted at, this is in most cases related to a memory issue. Can you ensure that memory availability can be ruled out?

1 Like