error: UnicodeDecodeError: 'utf-8' codec can't decode byte

[hongxianglics]

Hi,

I am trying to use the retrosynthetic transformer and I encountered some problem. I created a jupyter notebook in the folder askcos-core, and I can successfully import the packages with code
import time
import unittest
import askcos.global_config as gc
import askcos.retrosynthetic.transformer as retro_trans

however, when I try to use the transformer with the following code
t = retro_trans.RetroTransformer(load_all=False, use_db=True)
t.load()
I got the error
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xce in position 169: invalid continuation byte

the detailes can also be seen in the figures


I am confused about this and could you please help me with it? Many thansk

[connorcoley]

This looks like it might be a version issue with tensorflow, since it's failing to load the model (not elsewhere in the code); can you describe your environment?

[mliu49]

Another possibility is that you did not download the full model files when cloning askcos-data, which uses git-lfs to store large model files.

If you did not already have git-lfs installed when cloning, you may need to run git lfs pull in the askcos-data repo.