The [Speech2IPA Hugging Face Leaderboard](https://huggingface.co/spaces/KoelLabs/IPA-Transcription-EN) currently only supports Wav2Vec2ForCTC models from transformers. Would be useful to also [include Sherpa models](https://github.com/KoelLabs/ML/blob/main/scripts/ipa_transcription/sherpa.py).