wav2vec_test / tokenizer_config.json
othrif's picture
test of egyptian arabic with wav2vec2 fine-tuning
dee4b00
raw history blame
No virus
138 Bytes
{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|"}