viktor-enzell
/

wav2vec2-large-voxrex-swedish-4gram

Automatic Speech Recognition

hf-asr-leaderboard

Inference Endpoints

Model card Files Files and versions Community

wav2vec2-large-voxrex-swedish-4gram / language_model /attrs.json

viktor-enzell's picture

Added a 4-gram language model based on a 40M token social media corpus.

c877ba5 over 2 years ago

78 Bytes

{"alpha": 0.5, "beta": 1.5, "unk_score_offset": -10.0, "score_boundary": true}