Speech-Transformer | PyTorch implementation of Speech Transformer | Speech library

by kaituoxu Python Version: Current License: No License

X-Ray Key Features Code Snippets Community Discussions(1)Vulnerabilities Install Support

kandi X-RAY | Speech-Transformer Summary

Speech-Transformer is a Python library typically used in Artificial Intelligence, Speech, Pytorch, Transformer applications. Speech-Transformer has no bugs, it has no vulnerabilities, it has build file available and it has low support. You can download it from GitHub.

A PyTorch implementation of Speech Transformer [1], an end-to-end automatic speech recognition with Transformer network, which directly converts acoustic features to character sequence using a single nueral network.

Support

Quality

Security

License

Reuse

Support

Speech-Transformer has a low active ecosystem.

It has 709 star(s) with 194 fork(s). There are 31 watchers for this library.

It had no major release in the last 6 months.

There are 5 open issues and 34 have been closed. On average issues are closed in 73 days. There are no pull requests.

It has a neutral sentiment in the developer community.

The latest version of Speech-Transformer is current.

Quality

Speech-Transformer has 0 bugs and 35 code smells.

Security

Speech-Transformer has no vulnerabilities reported, and its dependent libraries have no vulnerabilities reported.

Speech-Transformer code analysis shows 0 unresolved vulnerabilities.

There are 3 security hotspots that need review.

License

Speech-Transformer does not have a standard license declared.

Check the repository for any license declaration and review the terms closely.

Without a license, all rights are reserved, and you cannot use the library in your applications.

Reuse

Speech-Transformer releases are not available. You will need to build from source code and install.

Build file is available. You can build the component from source.

Installation instructions, examples and code snippets are available.

Speech-Transformer saves you 568 person hours of effort in developing the same functionality from scratch.

It has 1327 lines of code, 63 functions and 26 files.

It has medium code complexity. Code complexity directly impacts maintainability of the code.

Top functions reviewed by kandi - BETA

kandi has reviewed Speech-Transformer and discovered the below as its top functions. This is intended to give you an instant insight into Speech-Transformer implemented functionality, and help decide if they suit your requirements.

Train the model
Run one epoch
Calculate the loss of the loss
Calculate the Calculation of the loss
Compute the decoder
Preprocess padded input
Get key - pad mask for key query
Pad a list
Compute the layer
Decodes a model
Build LFR features from inputs
Load a model from a package
Recognize beam
Extracts the sos_id and eos_id
Load a model
Reset checkpoint
Recognize the beam
Recognize a beam
Extracts the sos id and eos id
Load a model from a file

Get all kandi verified functions for this library.

Speech-Transformer Key Features

No Key Features are available at this moment for Speech-Transformer.

Speech-Transformer Examples and Code Snippets

No Code Snippets are available at this moment for Speech-Transformer.

Community Discussions

Trending Discussions on Speech-Transformer

What is attention penalty in speech transformer paper? (updated)

QUESTION

What is attention penalty in speech transformer paper? (updated)

Asked 2020-Jan-13 at 15:13

github: https://github.com/sephiroce/tfsr/tree/exprimental

I'm trying to reproduce recognition accuracies described in the speech transformer paper [1]. The attention penalty is a technique I could not fully understand. This is the description of the attention penalty in the paper.

"In addition, we encouraged the model attending to closer positions by adding bigger penalty on the attention weights of more distant position-pairs."

I understood as it means adding smaller negative values for more away from the diagonal on scaled attention logits (before masking) except for the first multi-head attention in decoders.

This is a code snippet for computing attention weights.

...

ANSWER

Answered 2020-Jan-13 at 10:33

I think you understand it well. They probably did a stripe around the diagonal, something like:

Source https://stackoverflow.com/questions/59646954

Community Discussions, Code Snippets contain sources that include Stack Exchange Network

Vulnerabilities

No vulnerabilities reported

Install Speech-Transformer

Python3 (recommend Anaconda)
PyTorch 0.4.1+
Kaldi (just for feature extraction)
pip install -r requirements.txt
cd tools; make KALDI=/path/to/kaldi
If you want to run egs/aishell/run.sh, download aishell dataset for free.
You can change parameter by $ bash run.sh --parameter_name parameter_value, egs, $ bash run.sh --stage 3. See parameter name in egs/aishell/run.sh before . utils/parse_options.sh.

Support

For any new features, suggestions and bugs create an issue on GitHub. If you have any questions check and ask questions on community page Stack Overflow .

Find more information at: