Datasets:

zszhong

/

Lyra-Data

like 2

Hi, thanks for your interest! The transcripts for the audio can be found in the following:

1. For the SFT part
   Lyra-Data/Lyra_SFT/multi_modality_speech/lyra_multimodal.json: key["speech_asr"]
   Lyra-Data/Lyra_SFT/long_speech/lyra_longspeech.json: key["speech_asr"]
2. For the pretraining part
   Lyra-Data/Lyra_Pretrain/lyra_pretrain.json: key["text"]

Sorry for the misunderstanding key name, we will rename it in our future version.