Giter VIP home page Giter VIP logo

kaldi-inspection's Introduction

MIR kaldi inspection tool

Introduction

把 prebuilt 好的 kaldi recipe (或基於 kaldi standard recipe 建立的 egs),將一個 decode 結果的 ctm (透過 steps/get_ctm_fast.sh 取得)render 成 html。

Image of demo listing Image of demo listing

Build steps

1. start server

  1. git clone https://github.com/kakushawn/kaldi-inspection

  2. cd kaldi-inspection

  3. pip install -r requirements.txt

  4. 修改 config.py,把 ROOT_FOLDER 改成對應路徑。

  5. 執行 python run.py 啟動 server,應該可以連到 localhost:5000

2. linking kaldi dependencies

  1. 將語料建立軟連結到 static/dataset/ 下。

    ln -s ~/kaldi-asr/egs/yesno/s5/waves_yesno static/dataset/

  2. 將 kaldi/scripts 腳本複製到 kaldi 腳本目錄下。

    cp kaldi/scripts/* ~/kaldi-asr/egs/aishell/s5/

  3. 在 kaldi 腳本目錄下,執行 gen_mir_web_ana_data.sh,把 decode 結果產生網站需要的文件。

  • usage:

    gen_mir_web_ana_data.sh corpus data_of_decode graph/lang decode_dir

    • corpus: 語料目錄名稱

    • data_of_decode: 該次 decode 執行時的 data 目錄

    • graph/lang: 該次 decode 時用到的 graph 或 lang 目錄

    • decode_dir: decode 完的目錄

    • example:

      cd ~/kaldi-asr/egs/yesno/s5/
      ./gen_mir_web_ana_data.sh waves_yesno data/test_yesno data/lang_test_tg exp/mono0a/decode_test_yesno/
      
  1. 回到網站目錄,在 kaldi/decodes/ 建立 decode 目錄的軟連結。

    ln -s ~/kaldi-asr/egs/yesno/s5/exp/mono0a/decode_test_yesno/ kaldi/decodes/

  2. 打開首頁,選項應該有剛剛新增的 decode 目錄。若要再新增其他目錄重複步驟 3 到 4 即可。

Notes

  1. 目前針對取得 audio resource 的方法非常 workaround,需要 flask server 與 corpus 在同一個 server,且只針對 wav 檔處理。

kaldi-inspection's People

Contributors

kakushawn avatar winedays avatar dependabot[bot] avatar

Stargazers

 avatar

Watchers

 avatar

kaldi-inspection's Issues

wav resource refinement

The way to locate a wav file heavily depends on position of corpus, which might not be robust enough.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.