broadinstitute / CellBender

CellBender is a software package for eliminating technical artifacts from high-throughput single-cell RNA sequencing (scRNA-seq) data.
https://cellbender.rtfd.io
BSD 3-Clause "New" or "Revised" License
280 stars 50 forks source link

Running CellBender without graphic card #161

Closed nadinegheziel closed 1 year ago

nadinegheziel commented 1 year ago

Hello, Thank you very much for your hard work. I am trying to run CellBender on my data, but I have an issue :

" assert torch.cuda.is_available(), "Trying to use CUDA, " \ AssertionError: Trying to use CUDA, but CUDA is not available."

I finally understood that I didn't have a graphic card on my work computer. Is it any way to run CellBender without graphic card ? I am very not familiar with all of this, i am sorry if it doesn't make sense.

Respectfully, Nadine

sjfleming commented 1 year ago

Hi Nadine,

Glad you're trying out CellBender!

There are two answers to your question:

  1. The simple answer is "yes": you can run CellBender without a GPU. You need to remove the flag --cuda from your cellbender command. So instead of
cellbender remove-background --cuda --input INPUT_FILE.h5 --output OUTPUT_FILE.h5 ...

(where the ... is the rest of the command: whatever other inputs you want to use) you would use the following command:

cellbender remove-background --input INPUT_FILE.h5 --output OUTPUT_FILE.h5 ...

(just omit --cuda)

  1. The slightly more complicated answer is that, while CellBender can run on CPU without a GPU, it is quite slow that way. CellBender is really meant to be run on a GPU. There are two options I know of for running CellBender on a GPU when you don't actually own a machine with a GPU:

Please let me know if you have any other questions as you try it out.

nadinegheziel commented 1 year ago

I can't thank you enough for your help, it all worked well and I could run CellBender on my data I will keep in touch once I will analyze the output

Thank you again, Kindly Nadine