Actually, you should be able to. I never really thought about it, but it is in the same format. You may need to try different bucket sizes. To run it, place it somewhere like:
data/raw/
You can then run:
python train.py --raw_data_dir="data/raw/"
I will need to update the documentation to explain this a bit better.
I've commited a dataset which will work. I'm closing this issue now, but you're welcome to re-open or start a new one if there are other problems/ if you have questions.
Actually, you should be able to. I never really thought about it, but it is in the same format. You may need to try different bucket sizes. To run it, place it somewhere like:
data/raw/
You can then run:
python train.py --raw_data_dir="data/raw/"
I will need to update the documentation to explain this a bit better.