flexudy/t5-small-wav2vec2-grammar-fixer

After transcribing your audio with Wav2Vec2, you might be interested in a post processor . All paragraphs had at most 128 tokens (separated by white spaces) all had at least 128 tokens . I strongly recommend improving the performance via further fine-tuning or by training more examples . Quick Rule based improvements: Align the transcribed version and the generated version . If the similarity of two words (case-insensitive) vary by more than some threshold based on some similarity metric (e.g. Levenshtein), then keep the transcriptions of the words that were transcribed and generated version. Align . Align your transcriptions with the correct version and,