Datasets:
Tasks:
Text-to-Speech
Modalities:
Audio
Formats:
soundfolder
Languages:
English
Size:
1K - 10K
License:
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
tags: [speech, text-to-speech, speaker-unlearning, voiceeraser]
|
| 4 |
+
---
|
| 5 |
+
# VoiceEraser — synthesized speech samples
|
| 6 |
+
|
| 7 |
+
Synthesized clips accompanying **VoiceEraser** (cond-DIM speaker unlearning, ICASSP 2026).
|
| 8 |
+
Each `*_clones/` folder holds, per speaker/sentence:
|
| 9 |
+
- `cB_*` — **baseline** clone from the pristine model,
|
| 10 |
+
- `cE_*` — **edited** clone after unlearning,
|
| 11 |
+
- `cRL_*` — clone after the white-box **relearn** attack.
|
| 12 |
+
|
| 13 |
+
Folders are named `<method>_<backbone>_clones/` (backbones: XTTS-v2, Tortoise, IndexTTS-1.5).
|
| 14 |
+
Audio is synthesized from the **consent-based** VCTK and LibriTTS research corpora and is provided
|
| 15 |
+
for **research/demonstration only**; it remains subject to those corpora's licenses.
|
| 16 |
+
|
| 17 |
+
Code: https://github.com/pujariaditya/voiceeraser
|