Unable to obtain testing accuracy as the tutorials

[hyang428]

Hi,

using the CMU-MOSI dataset:
the testing accuracy in the tutorial is around 71.6%.
I'm using the exactly same code from the tutorials, but I only get around 61.2% testing accuracy.

Any suggestion? or the author of CMU-MultimodalSDK has done something new to the dataset?

Thanks!

[Justin1904]

Not sure if this is still relevant, I'm guessing it could either be PyTorch updates or the SDK updates that made the difference. Maybe try something like PyTorch 1.0 to see if it gets ~70% acc, otherwise it's probably the SDK's data changed.

[bourne-3]

Hi,
I also encounter the same problem.I changed the model verified in the published paper but i keep the same implementation process in this repository.And the result shows a huge decrease.Maybe is the torch package updates some setting?Can you give some better implementation process in this tutorial?
Best wish