wer update

codes/data/audio/voice_tokenizer.py

@@ -13,13 +13,35 @@ from models.tacotron2.taco_utils import load_filepaths_and_text
 from models.tacotron2.text.cleaners import english_cleaners
 
 
+def remove_extraneous_punctuation(word):
+    replacement_punctuation = {
+        '{': '(', '}': ')',
+        '[': '(', ']': ')',
+        '`': '\'', '—': '-',
+        '—': '-', '`': '\'',
+        'ʼ': '\''
+    }
+    replace = re.compile("|".join([re.escape(k) for k in sorted(replacement_punctuation, key=len, reverse=True)]), flags=re.DOTALL)
+    word = replace.sub(lambda x: replacement_punctuation[x.group(0)], word)
+
+    # TODO: some of these are spoken ('@', '%', '+', etc). Integrate them into the cleaners.
+    extraneous = re.compile(r'^[@#%_=\$\^&\*\+\\]$')
+    word = extraneous.sub('', word)
+    return word
+
+
 class VoiceBpeTokenizer:
     def __init__(self, vocab_file):
-        self.tokenizer = Tokenizer.from_file(vocab_file)
+        if vocab_file is not None:
+            self.tokenizer = Tokenizer.from_file(vocab_file)
 
-    def encode(self, txt):
+    def preprocess_text(self, txt):
         txt = english_cleaners(txt)
         txt = remove_extraneous_punctuation(txt)
+        return txt
+
+    def encode(self, txt):
+        txt = self.preprocess_text(txt)
         txt = txt.replace(' ', '[SPACE]')
         return self.tokenizer.encode(txt).ids
 
@@ -28,6 +50,9 @@ class VoiceBpeTokenizer:
             seq = seq.cpu().numpy()
         txt = self.tokenizer.decode(seq, skip_special_tokens=False).replace(' ', '')
         txt = txt.replace('[SPACE]', ' ')
+        txt = txt.replace('[STOP]', '')
+        txt = txt.replace('[UNK]', '')
+
         return txt
 
 
@@ -50,23 +75,6 @@ def build_text_file_from_priors(priors, output):
             out.flush()
 
 
-def remove_extraneous_punctuation(word):
-    replacement_punctuation = {
-        '{': '(', '}': ')',
-        '[': '(', ']': ')',
-        '`': '\'', '—': '-',
-        '—': '-', '`': '\'',
-        'ʼ': '\''
-    }
-    replace = re.compile("|".join([re.escape(k) for k in sorted(replacement_punctuation, key=len, reverse=True)]), flags=re.DOTALL)
-    word = replace.sub(lambda x: replacement_punctuation[x.group(0)], word)
-
-    # TODO: some of these are spoken ('@', '%', '+', etc). Integrate them into the cleaners.
-    extraneous = re.compile(r'^[@#%_=\$\^&\*\+\\]$')
-    word = extraneous.sub('', word)
-    return word
-
-
 def train():
     with open('all_texts.txt', 'r', encoding='utf-8') as at:
         ttsd = at.readlines()

codes/scripts/audio/word_error_rate.py

@@ -5,6 +5,7 @@ import Levenshtein as Lev
 import torch
 from tqdm import tqdm
 
+from data.audio.voice_tokenizer import VoiceBpeTokenizer
 from models.tacotron2.text import cleaners
 
 
@@ -56,9 +57,27 @@ class WordErrorRate:
         return Lev.distance(''.join(w1), ''.join(w2))
 
 
+def load_truths(file):
+    niltok = VoiceBpeTokenizer(None)
+    out = {}
+    with open(file, 'r', encoding='utf-8') as f:
+        for line in f.readline():
+            spl = line.split('|')
+            if len(spl) != 2:
+                continue
+            path, truth = spl
+            path = path.replace('wav/', '')
+            truth = niltok.preprocess_text(truth)  # This may or may not be considered a "cheat", but the model is only trained on preprocessed text.
+            out[path] = truth
+    return out
+
+
 if __name__ == '__main__':
-    inference_tsv = 'D:\\dlas\\codes\\results.tsv'
-    libri_base = 'Z:\\libritts\\test-clean'
+    inference_tsv = 'results.tsv'
+    libri_base = '/h/bigasr_dataset/librispeech/test_clean/test_clean.txt'
+
+    # Pre-process truth values
+    truths = load_truths(libri_base)
 
     wer = WordErrorRate()
     wer_scores = []
@@ -67,13 +86,6 @@ if __name__ == '__main__':
         for line in tqdm(tsv):
             sentence_pred, wav = line.split('\t')
             sentence_pred = normalize_text(sentence_pred)
-
-            wav_comp = wav.split('_')
-            reader = wav_comp[0]
-            book = wav_comp[1]
-            txt_file = os.path.join(libri_base, reader, book, wav.replace('.wav', '.normalized.txt'))
-            with open(txt_file, 'r') as txt_file_hndl:
-                txt_uncleaned = txt_file_hndl.read()
-                sentence_real = normalize_text(clean_text(txt_uncleaned))
-                wer_scores.append(wer.wer_calc(sentence_real, sentence_pred))
+            sentence_real = normalize_text(truths[wav])
+            wer_scores.append(wer.wer_calc(sentence_real, sentence_pred))
     print(f"WER: {torch.tensor(wer_scores, dtype=torch.float).mean()}")