Model initialization step within colab using tpu and default configuration exits with error. about hypernerf HOT 7 OPEN

I got it to work by changing the line to: jnp.broadcast_to(jnp.array([last_sample_z]), z_vals[..., :1].shape)

from hypernerf.

saunair

Could you maybe share your version of the Colab notebook. I am still experiencing issues after changing the line to your suggestion... Thanks a lot for the help! :)

from hypernerf.

Hey @corlangerak, did you remove the pip install hypernerf line in the collab notebook? Because changing the line locally wouldn't reflect the change(You'll be loading the hypernerf pip install instead).
I suggest

1. Copying the notebook out of the notebook folder (basically into the project's root folder i.e. hypernerf/my_notebook.ipynb instead of hypernerf/notebooks/my_notebook.ipynb).
2. Apply the fix I suggested, and then run the notebook/collab.

from hypernerf.

@saunair Hi Nair, I faced the same issue. I tried to changed the line you suggested above though. Is it possible for you that giving me a little bit more specific instructions ?
Thank you in advance, Nair.

from hypernerf.

@hsauod check these two changes: ae29d1d#diff-433be35a4beb7eeee9224dcbe28ec97d53330cd175060905cd5217863674003cR114

and check the second cell in my notebook here: https://github.com/saunair/hypernerf/blob/main/notebooks/HyperNeRF_Training.ipynb

@corlangerak here you go. Sorry about the delay

from hypernerf.

@hsauod check these two changes: ae29d1d#diff-433be35a4beb7eeee9224dcbe28ec97d53330cd175060905cd5217863674003cR114

and check the second cell in my notebook here: https://github.com/saunair/hypernerf/blob/main/notebooks/HyperNeRF_Training.ipynb

@corlangerak here you go. Sorry about the delay

@saunair Hi Nair, thank you very much for your kindly explanation

from hypernerf.

Hi, incase anyone run into error, quick fix for colab is to install @saunair version of hypernerf

in cell 1 replace
!pip install git+https://github.com/google/hypernerf

with
!pip install git+https://github.com/saunair/hypernerf

Thanks again to @saunair for fixing error

from hypernerf.