Text-to-Speech
TensorBoard
ONNX
Piper
Welsh
English
piper-tts
tts
welsh
cymraeg
audio
accessibility
assistive-technology
screen-reader
Instructions to use techiaith/cy_GB-bu_tts with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Piper
How to use techiaith/cy_GB-bu_tts with Piper:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Fix dataset and quality fields in the voice config
Browse filesFix the voice metadata so Home Assistant and wyoming-piper show a proper name.
wyoming-piper builds the voice description from `dataset` and `audio.quality`, so these configs show up as "code (trainer)". It also registers `dataset` as an alias, so voices sharing `dataset: "code"` collide.
- `cy_GB-bu_tts.onnx.json`: `dataset` "code" → "bu_tts", `audio.quality` "trainer" → "medium"
Only these two fields change. The phoneme maps, the speaker maps and `phonemizer.data_version` are untouched, so synthesis output is identical.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_017bzA4Jzchn3xXbbUyztjXU
- cy_GB-bu_tts.onnx.json +2 -2
cy_GB-bu_tts.onnx.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"dataset": "
|
| 3 |
"audio": {
|
| 4 |
"sample_rate": 22050,
|
| 5 |
-
"quality": "
|
| 6 |
},
|
| 7 |
"espeak": {
|
| 8 |
"voice": "cy"
|
|
|
|
| 1 |
{
|
| 2 |
+
"dataset": "bu_tts",
|
| 3 |
"audio": {
|
| 4 |
"sample_rate": 22050,
|
| 5 |
+
"quality": "medium"
|
| 6 |
},
|
| 7 |
"espeak": {
|
| 8 |
"voice": "cy"
|