train.py -data data/demo -save_model demo-model
RuntimeError: Trying to backward through the graph a second time, but the buffers have already been freed. Specify retain_graph=True when calling backward the first time.
To run on pytorch 0.4, you must fix some errors because the lib hasn't been migrated to pytorch 0.4. I think if you are new, you should use pytorch 0.3.1 first.
You could try this fork. https://github.com/Ubiqus/OpenNMT-py
Hi all, sorry for the delay, but we just updated our code to support pytorch 0.4.
Closing this for now, please let us know if the current code doesn't work.
Most helpful comment
You could try this fork. https://github.com/Ubiqus/OpenNMT-py