This project implements a method for faster and more memory-efficient RNN-T computation, called `pruned rnnt`.
Note: There is also a fast RNNT loss implementation in [k2](https://github.com/k2-fsa/k2) project, which shares the same code here. We make `fast_rnnt` a stand-alone project in case someone wants only this rnnt loss.
## How does the pruned-rnnt work ?
We first obtain pruning bounds for the RNN-T recursion using a simple joiner network that is just an addition of the encoder and decoder, then we use those pruning bounds to evaluate the full, non-linear joiner network.
...
...
@@ -214,15 +216,15 @@ loss = fast_rnnt.rnnt_loss(
## Benchmarking
The [repo](https://github.com/csukuangfj/transducer-loss-benchmarking) compares the speed
and memory usage of several transducer losses, the summary in the following table is taken
from there, you can check the repository for more details.
The [repo](https://github.com/csukuangfj/transducer-loss-benchmarking) compares the speed and memory usage of several transducer losses, the summary in the following table is taken from there, you can check the repository for more details.
Note: As we declare above, `fast_rnnt` also implements in [k2](https://github.com/k2-fsa/k2) project, so `k2` and `fast_rnnt` are equivalent in the benchmarking.
|Name |Average step time (us) | Peak memory usage (MB)|
