ValueError: could not broadcast input array from shape (896) into shape (0)

Troubleshooting
1

Dear Cryosparc team,

I ran into error when start 2D with one of my dataset with no obvious reason. It’s K3 super-res movies and particles was extracted with 768 pixel and Fourier cropped to 192 or 128, neither of the size worked. I saws some similar error reports in the past but did not find obvious cause and solution to it. Thanks for your attention and help.

[CPU: 1.88 GB]   Traceback (most recent call last):
  File "cryosparc_worker/cryosparc_compute/run.py", line 84, in cryosparc_compute.run.main
  File "cryosparc_worker/cryosparc_compute/jobs/class2D/run.py", line 101, in cryosparc_compute.jobs.class2D.run.run_class_2D
  File "/gpfs0/home/jianshi/bin/cryosparc_worker/cryosparc_compute/sigproc.py", line 291, in generate_premultiplier
    sig[(Nsig//2)-(7*nsig//2):(Nsig//2)+(7*nsig//2)] = y
ValueError: could not broadcast input array from shape (896) into shape (0)
2

Hi @jamon, what do you have your 2D zeropad factor parameter set to in this job?

3

Yes. the default is 2, but there is cuda memory error, so I reduce it to 0 or 1

4

@jamon the zeropad should be greater than or equal to 1, then this should work. Can you share the full cuda memory error message you’re getting? What kind of GPU do you have?

5

@jamon do you have any updates?

6

How do you set the zeropad factor or see what it is? I am trying to downsample my particles and am getting a similar error:

ValueError: could not broadcast input array from shape (10,10) into shape (320,320)

7

I think I figured this one out- I was asking to downsample to a box size larger than what I previously had. Oops. Things seem to be running now.

2 Likes