mozilla/DeepSpeech

CC PMF importer improvements

Open

#3,450 建立於 2020年12月2日

在 GitHub 查看
 (2 留言) (0 反應) (0 負責人)C++ (26,755 star) (4,093 fork)batch import
good first bughelp wanted

描述

The current code does not work well:

  • extraction fails because Python zip does not support multiarchive, we should rely on 7z for example
  • we should expose MIN_SECS / MAX_SECS to the command-line arguments
  • document ffmpeg dependency for audio conversion
  • we use multiprocess for performing one WAV split, but maybe we should also use multiprocess for converting multiple MP3 to WAV in paralell

貢獻者指南

技術棧
python
領域
datamachine learningtooling
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-2 days
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
Pythonbasic understanding of audio processing
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
40
研究方向
The issue involves improving the Common Voice dataset importer. First, review the current extraction logic and replace Python's zip with 7z to handle multi archive files. Then, add command line arguments for MIN SECS and MAX SECS to allow user configuration. Document the ffmpeg dependency needed for audio conversion. Finally, explore using multiprocessing to parallelize MP3 to WAV conversion. Check the repository's importer module for existing code.