-
Notifications
You must be signed in to change notification settings - Fork 80
Description
I was running cellbender remove-background on my data and this is the code i used.
cellbender remove-background --cuda --input WT.4.h5 --output WT.4_cb.h5
The actual issue is this :
File "/home/athirah/miniconda3/envs/cellbender/lib/python3.7/site-packages/cellbender/remove_background/vae/encoder.py", line 318, in forward
assert cells.sum() > 4, "Fewer than 4 cells passed to encoder minibatch"
AssertionError: Fewer than 4 cells passed to encoder minibatch
I tried different values of --empty-droplet-training-fractionand adding --expected-cells that too didnt work.
Environment
GPU
Model: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
Compute capability: sm_120
Driver version: 580.65.06
CUDA version (driver): 13.0
CUDA Toolkit
nvcc version: 12.0 (V12.0.140)
Python version: 3.7
PyTorch version: 1.13.1+cu117
PyTorch CUDA version: 11.7
cellbender 0.3.0
Please look into this.