Pre-trained model

[remega]

Hi Alami,

Thanks for the nice work!

I am trying to run your project with my own data. Following your introduction, when I was trying to frozen model, it appears error with the default output_node_names (here). Could provide this setting to me that I can directly use your pre-trained model?

Besides, I found there was a frozen model (frozen_model.pb) in the fold of "increase" . However, when I was trying this in infer_HR.py, it also appears shape-unmatched error:
InvalidArgumentError (see above for traceback): Incompatible shapes: [1,156,256,3] vs. [1,311,512,1]
[[node sharpening_station/mul_4 (defined at /home/remega/projects/Gaze-Shift-Net/utils.py:56) ]]
[[node clip_by_value (defined at /home/remega/projects/Gaze-Shift-Net/utils.py:56) ]]
So could you please provide the right frozen_model.pb that I can directly run the infer_HR.py? Or anything I can do for that?

I would appreciate much if any advice.

Thanks,
Lai