Speech output is taking infinity time

[Ashbajawed]

T->T and S->T aer working fine but tasks with speech outputs like T->S and S->S are taking infinite time

Also I would like to know how can we utilize this model for ASR and TTS tasks

[Sunbread]

this model might be a base model, with no instruction sft. you should fine-tune it before asr or tts tasks

[hitchhicker]

Hey, you could either further fine-tune the model with ASR/TTS instructions or use few-shot examples to guide the model for ASR/TTS, the latter option is covered in the appendix B of the paper.

I don't understand what do you mean by taking infinite time though. Could you give concrete example so that we can reproduce? Thanks!