* Disable weight decay.
* Remove input feature batchnorm..
* Replace BatchNorm in the Conformer model with LayerNorm.
* Use tanh in the joint network.
* Remove sos ID.
* Reduce the number of decoder layers from 4 to 2.
* Minor fixes.
* Fix typos.
* Begin to add RNN-T training for librispeech.
* Copy files from conformer_ctc.
Will edit it.
* Use conformer/transformer model as encoder.
* Begin to add training script.
* Add training code.
* Remove long utterances to avoid OOM when a large max_duraiton is used.
* Begin to add decoding script.
* Add decoding script.
* Minor fixes.
* Add beam search.
* Use LSTM layers for the encoder.
Need more tunings.
* Use stateless decoder.
* Minor fixes to make it ready for merge.
* Fix README.
* Update RESULT.md to include RNN-T Conformer.
* Minor fixes.
* Fix tests.
* Minor fixes.
* Minor fixes.
* Fix tests.
* Apply layer normalization to the output of each gate in LSTM.
* Apply layer normalization to the output of each gate in GRU.
* Add projection support to LayerNormLSTMCell.
* Add GPU tests.
* Use typeguard.check_argument_types() to validate type annotations.
* Add typeguard as a requirement.
* Minor fixes.
* Fix CI.
* Fix CI.
* Fix test failures for torch 1.8.0
* Fix errors.
* add MMI to AIShell
* fix MMI decode graph
* export model
* typo
* fix code style
* typo
* fix data prepare to just use train text by uid
* use a faster way to get the intersection of train and aishell_transcript_v0.8.txt
* update AIShell result
* update
* typo
We are using multiple machines to do various experiments. It makes
life easier to know which experiment is running on which machine
if we also log the IP and hostname of the machine.
* Modify label smoothing to match the one implemented in PyTorch.
* Enable CI for torch 1.10
* Fix CI errors.
* Fix CI installation errors.
* Fix CI installation errors.
* Minor fixes.
* Minor fixes.
* Minor fixes.
* Minor fixes.
* Minor fixes.
* Fix CI errors.
* Update RESULTS using vocab size 500, att rate 0.8
* Update README.
* Refactoring.
Since FSAs in an Nbest object are linear in structure, we can
add the scores of a path to compute the total scores.
* Update documentation.
* Change default vocab size from 5000 to 500.
* Add a note about the CUDA OOM error.
Some users consider this kind of OOM as an error during decoding,
but actually it is not. This pull request clarifies that.
* Fix style issues.
* add a docker file for some users
Ubuntu18.04-pytorch1.7.1-cuda11.0-cudnn8-python3.8
* add a describing file of how to use dockerfile
give some steps to use dockerfile
