elgeish/wav2vec2-base-timit-asr

The model can be used directly (without a language model) as follows . Make sure that your speech input is sampled at 16kHz . You find the fine-tuning script used to produce this output . The model is based on the timit_asr dataset . It is used to generate the output of the soundfile soundfile and other datasets . For example, make sure that the sound file is soundfile is sound file with soundfile as sf. The model was used without a language or language model, so it can be easily used directly without the help of the language model. The output is produced using the Wav2Vec2ForCTC, WavVec,