Igor Tatarnikov
Igor Tatarnikov
It seems there are some issues with the obj files for the Kim atlas. I sampled a few other random areas and there's a high chance that the obj files...
I would agree, it seems to be an issue with the generation. We should transfer it to `bg-atlasgen`!
Regenerating the meshes seems to fix this issue. 
I think this may be an underlying issue with the atlas meshes themselves, see [brainglobe/brainglobe-atlasapi#236](https://github.com/brainglobe/brainglobe-atlasapi/issues/236)
Fixed in #281  
Success! I was able to replicate the failure on an Ubuntu 22.04 VM with 4 GB of RAM and 4 cores. More clues in the full trace: ``` ================================================================== FAILURES...
Seems the error itself is sporadic! ``` _______________________________________________________________ test_callbacks ________________________________________________________________ signal_array = dask.array background_array = dask.array, no_free_cpus = 0 def test_callbacks(signal_array, background_array, no_free_cpus): # 20 is minimum number of planes...
Interestingly, I couldn't reproduce the failure no matter how I played with the n_free_cpus parameter. Could only reproduce in the VM. Setting the tests to run with 1 free CPU...
It seems to lie at the intersection of low core count machine with JIT compilation disabled in `numba`. If I set the VM core count to 8 I can no...
Numbers from the paper, original cellfinder with a tensorflow backend, default model (lighter points) vs retrained model (darker points)  `cellfinder`, torch backend, default model (lighter points) vs retrained model...