Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Support different tts model types. #1541

Merged
merged 7 commits into from
Mar 12, 2024
Merged

Conversation

csukuangfj
Copy link
Collaborator

low (quality) -> runs faster. high (quality) -> runs slower


Need to test it before merging.

low (quality) -> runs faster. high (quality) -> runs slower
@@ -38,6 +39,36 @@
"hifigan_multi_scale_multi_period_discriminator": HiFiGANMultiScaleMultiPeriodDiscriminator, # NOQA
}

LOW_CONFIG = {
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

quantize_dynamic(
model_input=model_filename,
model_output=model_filename_int8,
weight_type=QuantType.QUInt8,
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Quantizing using quint8 is very slow at run time, so we removed it.

@csukuangfj
Copy link
Collaborator Author

csukuangfj commented Mar 12, 2024

The following wave is generated from a model trained with --model-type low.

low.mp4

The following wave is generated with a model trained with --model-type medium

medium.mp4

The following wave is generated with a model trained with --model-type high

high.mp4

@csukuangfj csukuangfj merged commit 81f518e into k2-fsa:master Mar 12, 2024
64 checks passed
@csukuangfj csukuangfj deleted the tts-fix branch March 12, 2024 14:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant