mLSTM with factored data
Reported on 1 benchmark across 1 task
Note: results are matched by exact model name. Different papers may use the same name for different model variants.
Natural Language Processing1 result
- 20.8best: 24.37 (Transformer trained on highly filtered data)