speechT

speetcht

用TensorFlow編寫的OpenSource語音到文本軟件。在Librispeech測試語料庫中，達到8％的字母錯誤率，單詞錯誤率為20％ 。

安裝

先決條件

需要Python3，Portaudio19-Dev和FFMPEG。

在Ubuntu安裝上

 sudo apt install python3-pip portaudio19-dev ffmpeg

通過PIP3安裝

 pip3 install git+https://github.com/timediv/speechT

建築學

當前的SpeechT基於WAV2LETTER紙張和CTC損耗函數。

來自http://www.opensl.org/12/的語音語料庫自動下載。
注意：語料庫約為30GB！

訓練

數據必須在培訓之前進行預處理

 speecht-cli preprocess

然後，進行培訓，執行

 speecht-cli train

使用--help以獲取更多詳細信息。

您可以監視培訓並在張板中查看其他日誌

 tensorboard --logdir log/

測試

在整個測試集運行中評估

 speecht-cli evaluate

評估一批

 speecht-cli evaluate --step-count 1

默認情況下，使用貪婪解碼。請參閱有關如何使用Kenlm進行解碼的Using a language model部分。

使用--help以獲取更多詳細信息。

實時用法

使用麥克風記錄並打印預測運行

 speecht-cli record

使用--help以獲取更多詳細信息。

訓練有素的重量

您沒有自己的資源可以自己訓練？從這裡下載權重

 mkdir train
tar xf speechT-weights.tgz -C train/

然後，您可以將模型與EG評估一起使用

 speecht-cli evaluate --run-name best_run

使用語言模型

如果您想將KENLM用作解碼的語言模型，則需要編譯並安裝tensorflow-with-kenlm。如果您僅需要Linux的CPU版本的TensorFlow版本，則也可以在此處下載。

從這裡下載所有必要的文件，然後

 tar xf kenlm-english.tgz
speecht-cli evaluate --language-model kenlm-english/

結果

默認參數在NVIDIA TITAN X上進行了約5至6天的訓練。

總體統計數據

 Average Letter Edit Distance: 7.7125
Average Letter Error Rate: 8%
Average Word Edit Distance: 3.801953125
Average Word Error Rate: 20%

在一些示例中進行的預測和預測

 expected: but that is kaffar's knife
decoded: but that is caffr's klife 
LED: 4 LER: 0.15 WED: 2 WER: 0.40

expected: he moved uneasily and his chair creaked
decoded: he moved uneasily in his chair creet
LED: 5 LER: 0.13 WED: 2 WER: 0.29

expected: it is indeed true that the importance of tact and skill in the training of the young and of cultivating their reason and securing their affection can not be overrated
decoded: it is indeed true that the importance of tact and skill in the training of the young and of cultivating their reason and so carrying their affection can not be o rated
LED: 8 LER: 0.05 WED: 4 WER: 0.13

expected: she pressed his hand gently in gratitude
decoded: she pressed his hand gently in gratitude
LED: 0 LER: 0.00 WED: 0 WER: 0.00

expected: don't worry sizzle dear it'll all come right pretty soon
decoded: don't worry i l dear it all come riprety soon 
LED: 13 LER: 0.23 WED: 5 WER: 0.50

expected: may we see gates at once asked kenneth
decoded: may we see gates at once asked keneth 
LED: 2 LER: 0.05 WED: 1 WER: 0.12

可以在此處找到整個評估日誌。