v0.0.2

Changes

• Lexicon validator in PER algorithm
• Generalize Transcriber to any-level offsets (word, character/phoneme, sentence)
• Refactored Submodules
• Extend customizability of Runner
• Implemented Word-overlap Segmentation
• Removed batched audio classification
• Output dataset to HuggingFace Dataset utility script
• Added guides, demos, examples

v0.0.1

First publication of SpeechLine, supports basic audio classification, transcription, and segmentation.