Real speech for the audio example: LibriSpeech dev-clean 1462-170145-0022
Browse files
README.md
CHANGED
|
@@ -53,6 +53,14 @@ Rules the model imposes, enforced here before anything is uploaded:
|
|
| 53 |
which is why the duration slider disappears when a single reference can set it, and comes back when two can or
|
| 54 |
when the one that could is out of range.
|
| 55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
## How the split is expressed
|
| 57 |
|
| 58 |
`MiniMaxH3Ref2VABlocks` is a `SequentialPipelineBlocks` of eight steps:
|
|
|
|
| 53 |
which is why the duration slider disappears when a single reference can set it, and comes back when two can or
|
| 54 |
when the one that could is out of range.
|
| 55 |
|
| 56 |
+
## Example assets
|
| 57 |
+
|
| 58 |
+
`examples/subject.png` is generated (FLUX.1-schnell), `examples/motion.mp4` is a synthetic clip from the parity
|
| 59 |
+
fixtures, and `examples/voice.wav` is utterance `1462-170145-0022` of
|
| 60 |
+
[LibriSpeech](https://www.openslr.org/12) `dev-clean` — CC BY 4.0, read from a public-domain LibriVox recording. It
|
| 61 |
+
is 16 kHz mono on purpose: the audio VAE wants 32 kHz, so the example exercises the `torchaudio` resample the
|
| 62 |
+
`ref2va` path needs.
|
| 63 |
+
|
| 64 |
## How the split is expressed
|
| 65 |
|
| 66 |
`MiniMaxH3Ref2VABlocks` is a `SequentialPipelineBlocks` of eight steps:
|