I am currently benchmarking my own methods on the ADE corpus and I would like the results to be directly comparable to your paper ("A neural joint model for entity and relation extraction from biomedical text").
Are you able to share your splits? Most importantly, which 10% of examples did you hold out for testing?
Hello,
I am currently benchmarking my own methods on the ADE corpus and I would like the results to be directly comparable to your paper ("A neural joint model for entity and relation extraction from biomedical text").
Are you able to share your splits? Most importantly, which 10% of examples did you hold out for testing?
Thanks!