elbayadm/attn2d
Pervasive Attention: 2D Convolutional Networks for Sequence-to-Sequence Prediction
Implements 2D convolutions that jointly encode source and target sequences for neural machine translation, enabling efficient decoding grids for wait-k simultaneous translation. Extends Fairseq with Transformer-based wait-k models featuring unidirectional encoders and joint multi-path training. Built on PyTorch with GPU acceleration via NCCL for distributed training.
498 stars. No commits in the last 6 months.
Stars
498
Forks
71
Language
Python
License
MIT
Category
Last pushed
May 08, 2021
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/elbayadm/attn2d"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
luozhouyang/transformers-keras
Transformer-based models implemented in tensorflow 2.x(using keras).
xv44586/toolkit4nlp
transformers implement (architecture, task example, serving and more)
ufal/neuralmonkey
An open-source tool for sequence learning in NLP built on TensorFlow.
uzaymacar/attention-mechanisms
Implementations for a family of attention mechanisms, suitable for all kinds of natural language...
graykode/xlnet-Pytorch
Simple XLNet implementation with Pytorch Wrapper