search

burchim / EfficientConformer

[ASRU 2021] Efficient Conformer: Progressive Downsampling and Grouped Attention for Automatic Speech Recognition
https://arxiv.org/abs/2109.01163
Apache License 2.0
210 stars 32 forks source link

Could you share your id of the "LibriSpeech_bpe_1000.model" #22

Closed ZhengWenrui closed 1 year ago

ZhengWenrui commented 1 year ago

Hello, i find the id of the "LibriSpeech_bpe_256.model", but i want to download the "LibriSpeech_bpe_1000.model" for the training of Transducer, can you tell me the id of the "LibriSpeech_bpe_1000.model"? Thank you very much!

ZhengWenrui commented 1 year ago

Thanks, i may know how to create a BPE model. Ii can use "--create_tokenizer" to create a BPE model!

burchim commented 1 year ago

Hi, I do not possess the exact 1000 bpe tokenizer anymore but yes you can simply use the "--create_tokenizer" flag to automatically train a tokenizer for the Transducer model.

ZhengWenrui commented 1 year ago

Thank you very much!