Nik Afiq
5238298b55
- Implemented core model components in `modules.py` including various convolutional layers and normalization techniques. - Added transformation functions in `transforms.py` for piecewise rational quadratic transformations. - Created utility functions in `utils.py` for checkpoint management, logging, and hyperparameter handling. - Introduced monotonic alignment functionality with Cython optimization in `monotonic_align`. - Developed a minimal inference server in `server.py` to handle synthesis requests. - Updated requirements to include necessary dependencies for Cython and scipy.
33 lines
947 B
Protocol Buffer
33 lines
947 B
Protocol Buffer
syntax = "proto3";
|
|
package tts.v1;
|
|
option go_package = "gitea.nik4nao.com/nik/home-services/gen/tts/v1;ttsv1";
|
|
|
|
service TTSService {
|
|
rpc Synthesize(SynthesizeRequest) returns (SynthesizeResponse);
|
|
rpc ListSpeakers(ListSpeakersRequest) returns (ListSpeakersResponse);
|
|
}
|
|
|
|
// optional fields let the gateway apply its own defaults (matching the
|
|
// reference uma-tts-api's noise_scale=0.37, noise_scale_w=0.46,
|
|
// length_scale=1.3) when a client omits them.
|
|
message SynthesizeRequest {
|
|
string speaker_name = 1;
|
|
string text = 2;
|
|
optional float noise_scale = 3;
|
|
optional float noise_scale_w = 4;
|
|
optional float length_scale = 5;
|
|
}
|
|
|
|
message SynthesizeResponse {
|
|
bytes audio = 1;
|
|
string mime_type = 2;
|
|
}
|
|
|
|
message ListSpeakersRequest {
|
|
string search = 1; // case-insensitive substring filter; empty returns all speakers
|
|
}
|
|
|
|
message ListSpeakersResponse {
|
|
repeated string speaker_names = 1;
|
|
}
|