TotalSegmentator error at first run: ‘DummyFile’ object has no attribute ‘flush’

Hi Iassoan, I’m trying to use the TotalSegmentator Tool but i receive this error below the Apply button:

RuntimeError: FIND was unable to find an engine to execute this computation

Exception ignored in: <totalsegmentator.libs.DummyFile object at 0x000002765B9EE4F0>

AttributeError: ‘DummyFile’ object has no attribute ‘flush’

I checked if could be an error due to the GPU’s memory but on my laptop is 10Gb of Memory.
Thanks for you time

Could you provide some more details? What is the brand and model of the GPU?

Sure. The model is “GPU intel(r) uhd graphics 620”

Unfortunately, GPU processing will require an NVIDIA GPU, so you could switch to CPU processing which will take longer to run (30 minutes vs. 2 minutes).

Switch here:


I don’t know why but it doesn’t switch the computation on CPU, it still remain on GPU and ask me if i want to enable the ‘fast’ mode.

You need to uninstall the GPU version of pytorch, restart Slicer, and install the CPU version of pytorch - as described

1 Like

Ok, now it is correctly working. Thank you so much for your time

1 Like