Leandro Gracia Gil

Hi @teijeong, I'm trying to export a TFLite model with 2 signatures, now that multiple of them are supported in tf-nightly. One signature uses integer quantization and takes quantized inputs....

I'd be interested on getting this too. I'm looking for an op to do least squares (torch.linalg.lstsq), but it should be possible to do it through a SVD op and...

Note: this example is for Japanese, but I expect to do the same (feeding phonemes as input) in a few other latin script languages. I haven't checked yet if these...

Also, one thing I had to fix while debugging. I can open a separate bug if needed. In file tokenization/japanese.py, line 19: ```python config_path = resource_dir.joinpath("japanese", "sudachi_config.json") ``` This fails...

I see, thanks. Regardless from the tokenization issue, is there any other new feature, or improvement in quality I would be missing by using the old 2.0.1a model instead of...

For additional context, all utterances in the examples above should have a duration between 8 and 30 seconds.

Maybe it would be worth reporting this also as an issue to the [Dreambooth extension repo](https://github.com/d8ahazard/sd_dreambooth_extension/issues), linking back here.

Note: this should probably have a very low priority since there's a quite simple workaround: ```python def shape_n(input, out_type=tf.int32): return [tf.cast(tf.shape(t), out_type) for t in input] ``` That's not exactly...

> @leandro-gracia-gil Did you figure out how the mel files should look like? There is not much detail into it. No, at the end I changed meldataset.py locally to generate...

I'm closing this bug as the previous 2.0.1a model can still be used in this way. Thanks for your help.