Streaming infer

[wntg]

I want to use wavtokenizer to speech AI. Is avtokenizer apply streaming infer?

[jishengpeng]

I want to use wavtokenizer to speech AI. Is avtokenizer apply streaming infer?

Since WavTokenizer follows the same paradigm as Encodec and DAC, it can be adapted for streaming. However, this requires retraining with causal convolution. Based on empirical evidence, this retraining is unlikely to result in performance degradation.

[handsomelys]

I want to use wavtokenizer to speech AI. Is avtokenizer apply streaming infer?

Since WavTokenizer follows the same paradigm as Encodec and DAC, it can be adapted for streaming. However, this requires retraining with causal convolution. Based on empirical evidence, this retraining is unlikely to result in performance degradation.

Do I need to replace all convolutional layers involved in the WavTokenizer network with casual convolution?

[jishengpeng]

I want to use wavtokenizer to speech AI. Is avtokenizer apply streaming infer?

Since WavTokenizer follows the same paradigm as Encodec and DAC, it can be adapted for streaming. However, this requires retraining with causal convolution. Based on empirical evidence, this retraining is unlikely to result in performance degradation.

Do I need to replace all convolutional layers involved in the WavTokenizer network with casual convolution?

yes, and attention part.