Open zhf3564859793 opened 7 months ago
Hello, thanks for asking!
You do not necessarily need that .pkl
file, as it's simply a cache of the data in the ./data
folder for faster processing. As long as you have source data files there, be it our data for illustration or your own data, you can use the preprocess
function in the TAGConv/tagconv_baseline/src/data_processing.py
to generate the .pkl
file. I know the TAGConv/tagconv_quantum/src/data_processing_cache.py
is probably not as well-written as it can be, so you might need to modify it a little bit. You can refer to TAGConv/tagconv_baseline/src/data_processing.py
to modify as appropriate.
I would like to ask if you could provide the pickle document(ECFP_split_feature.pkl) in TAGConv/tagconv_quantum/src/data_processing_cache.py(line 126: pkl_path = '/home/lichangyu/Mercury/USPTO/data/ECFP_split/ECFP_split_feature.pkl'). I really appreciate your nice help!