Datasets:
Missing ['id'] Attribute in train_chime4.pt
#2
by
yizirui
- opened
Hello, and thank you for sharing your impressive work along with the code, model, and data. I've been following your provided instructions to load the train_chime4.pt file using add_speech_feats_to_train_data.py. However, it appears that each item in the data does not include the ['id'] attribute as expected. Could you provide any guidance on how to resolve this? Thank you!
Hi, thanks for report. 'id' is utterance id which is same as the file name (without '.wav') for chime4.