Kaizhi Qian

Results 196 comments of Kaizhi Qian

Follow-up work available for viewing

@qq547276542 Thanks! More follow-up works will be released. Stay tuned.

Issues with conversion of VCTK speakers using pre-trained model

The pretrained model is only trained on a small set of speakers, which may not generalize well to other speakers. You can use one-hot embedding if you are not doing...

The validation loss is rising and fluctuating, is that a regular situation?

The provided training data is very small for code verification purposes only.

The validation loss is rising and fluctuating, is that a regular situation?

There might be something wrong with your validation data. The validation loss should be around 30.

The validation loss is rising and fluctuating, is that a regular situation?

The format is correct.

The validation loss is rising and fluctuating, is that a regular situation?

@c1a1o1 Please clearly state your question and create a new issue. Please do NOT flood other issues.

The validation loss is rising and fluctuating, is that a regular situation?

@jamesliu This looks like over-fitting to me. Make sure you use a large training set and the validation speakers are in the training set.

The validation loss is rising and fluctuating, is that a regular situation?

@jamesliu Your training set is actually very small, which has only 30 mins of data. Also, the "demo data" needs to be consistent with the training data.

The validation loss is rising and fluctuating, is that a regular situation?

There are many contributing factors to output quality. It is hard to tell from the information you provided. @jamesliu

The validation loss is rising and fluctuating, is that a regular situation?

You can use longer audio. There is no limit on the length of input.