Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
@@ -6,22 +6,44 @@ colorTo: indigo
|
|
6 |
sdk: static
|
7 |
pinned: false
|
8 |
license: apache-2.0
|
9 |
-
short_description: My First
|
10 |
---
|
11 |
|
12 |
-
#
|
13 |
|
14 |
-
|
15 |
|
16 |
-
|
17 |
-
```
|
18 |
-
@article{park2021nerfies
|
19 |
-
author = {Park, Keunhong and Sinha, Utkarsh and Barron, Jonathan T. and Bouaziz, Sofien and Goldman, Dan B and Seitz, Steven M. and Martin-Brualla, Ricardo},
|
20 |
-
title = {Nerfies: Deformable Neural Radiance Fields},
|
21 |
-
journal = {ICCV},
|
22 |
-
year = {2021},
|
23 |
-
}
|
24 |
-
```
|
25 |
|
26 |
-
|
27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
sdk: static
|
7 |
pinned: false
|
8 |
license: apache-2.0
|
9 |
+
short_description: My First Binary Tree into AI
|
10 |
---
|
11 |
|
12 |
+
# GHOSTVOICECBR
|
13 |
|
14 |
+
**GHOSTVOICECBR** is a real-time voice cloning framework built using a novel Case-Based Reasoning (CBR) binary tree and quadrant-based tone mapping system. It enables low-latency, agentic voice synthesis using emotional quads β pitch, timbre, speed, and mood β mapped onto cloned neural speech vectors.
|
15 |
|
16 |
+
This system is the voice synthesis engine behind the **GhostVoice AI project**, capable of generating cloned speech in real time, optimized for 8β12 GB VRAM and deployable via Hugging Face Spaces, Gradio, or Twitch integrations.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
|
18 |
+
---
|
19 |
+
|
20 |
+
## β¨ Features
|
21 |
+
|
22 |
+
- β
**CBR Tree Traversal** for selecting voice tones over time
|
23 |
+
- β
**Quad Mapping Engine** for emotion-driven synthesis
|
24 |
+
- β
**Fast Inference** on local hardware (no cloud GPU required)
|
25 |
+
- β
**Speaker Embedding Support** for personalized cloning
|
26 |
+
- β
**Live Gradio UI** with console, waveform, and quad controls
|
27 |
+
- β
**Twitch VoiceBot Ready** (optional module)
|
28 |
+
|
29 |
+
---
|
30 |
+
|
31 |
+
## π§ Core Concepts
|
32 |
+
|
33 |
+
- **CBR Binary Tree**: Stores and retrieves historical tone vectors efficiently.
|
34 |
+
- **Quad Mapping**: Each speech sample is mapped using a 4D vector:
|
35 |
+
- `pitch`
|
36 |
+
- `speed`
|
37 |
+
- `timbre`
|
38 |
+
- `emotion`
|
39 |
+
- **Voice Matching**: Nearest neighbor match + synthetic generation
|
40 |
+
- **Open Format**: Easily extensible to other TTS models or APIs (Bark, MusicGen, etc.)
|
41 |
+
|
42 |
+
---
|
43 |
+
|
44 |
+
## π How to Use
|
45 |
+
|
46 |
+
1. Clone or fork this repo
|
47 |
+
2. Install dependencies:
|
48 |
+
```bash
|
49 |
+
pip install -r requirements.txt
|