OSError: Unable to open file (file signature not found)

[tashrifbillah]

Hi Vishnu (@vishnubashyam), did you upload the proper Models/DBN_model.h5 file to GitHub? I have been helping one of our RAs to install it and Python complained as follows when we tried Script/test.sh:

  File "py3-akb67/envs/akb67-deepbrain/lib/python3.6/site-packages/h5py/_hl/files.py", line 190, in make_fid
    fid = h5f.open(name, flags, fapl=fapl)
  File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
  File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
  File "h5py/h5f.pyx", line 96, in h5py.h5f.open
OSError: Unable to open file (file signature not found)

In my experience, h5 files are big and are circulated through GitHub release or Google Drive share.

[DVSneuro]

Hi @tashrifbillah -- not sure if you ever stumbled across a solution here, but I think the h5 files are on Box (https://upenn.app.box.com/v/DeepBrainNet). However, I still get the error below when using any of those h5 files:

ValueError: Layer count mismatch when loading weights from file. Model expected 1 layers, found 3 saved layers.

[tashrifbillah]

The RA left our office. So I never got to work on it anymore. Sorry!

[DVSneuro]

No worries! If you ever get back to it, @tannerjared has a nice fork of this repo with some fixes and more documentation (https://github.com/tannerjared/DeepBrainNet).

I still got the ValueError above, but it appears to be resolved by downgrading to tensorflow==2.15.

[tannerjared]

That's good feedback @DVSneuro about the working version of tensorflow. I want to add that @mateuszpawlik has a containerized version of this implemented in Nextflow

I haven't finished comparisons yet between this version and the containerized versions, but it's worth a look. It's a lovely, much more automated approach.

[DVSneuro]

Good to know, thanks!