Reproduce the results from the IHP paper

Hey @ManuelLobo @dpavot @fjmc ,

I am trying to reproduce the results from the IHP paper using the dockerfile with no luck.
Can you please help with the right commands needed to build the model and produce those numbers?

Thanks

[dpavot]

Hey @omriFdna,

Are you having trouble building the docker container and running the README.md commands or achieving the exact results stated in the IHP paper? If it is the latest, we no longer have access to the original models because we are in the process of creating a new version of IHP, but we contacted the first author asking if he saved a copy of the original models.
If the trouble is in the building/running stage can you be more specific about the problem?

Best

Thanks @dpavot for answering.
The build and README commands works great.
I am trying to use them to train the model described in the paper and use the evaluate.py in order to achieve the 0.863 F-score reported in the paper on the GSC+ set.
Is this possible?

[dpavot]

We contacted the first author who no longer works in our lab to see if he can provide more detailed guidelines to reproduce the numbers reported on the paper. Once we have an answer, we will post it here.