OldestSalt commited on
Commit
d6e1a51
·
verified ·
1 Parent(s): 6cbe3a7

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -19,7 +19,7 @@ Bimba is *almost* linear SimulMT model trained with wait-k policy (k = 3, 5, 7,
19
 
20
  The model has encoder-decoder architecture, where self-attention blocks are Mamba-2 blocks instead. It means that encoder is linear, but cross-attention's input is all outputs of encoder, and this means that complexity of Bimba is O(S * T), which is not *exactly* linear
21
 
22
- Bimba was developed as a part of master's thesis, and I hope that I can continue research in Linear SimulMT field/
23
 
24
  - Code: https://github.com/OldestSalt/LinearSimultMT
25
  - Paper: Soon
 
19
 
20
  The model has encoder-decoder architecture, where self-attention blocks are Mamba-2 blocks instead. It means that encoder is linear, but cross-attention's input is all outputs of encoder, and this means that complexity of Bimba is O(S * T), which is not *exactly* linear
21
 
22
+ Bimba was developed as a part of master's thesis, and I hope that I can continue research in Linear SimulMT field.
23
 
24
  - Code: https://github.com/OldestSalt/LinearSimultMT
25
  - Paper: Soon