File size: 1,813 Bytes
495c871 47afb5e 495c871 f4a9d34 495c871 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 | ---
license: apache-2.0
tags:
- voice-changer
- rvc
- spin
- f0-categorized
- non-speech
- pretrained-model
datasets:
- custom
language:
- multilingual
---
# 🔊 KLM BlackTone Large (Pretrained Voice Changer Model)
**KLM BlackTone Large** is a pretrained model for real-time **voice conversion**, built using the **Spin Embedder**. It was trained on a massive dataset of over **11,000 hours** from **595 speakers**, carefully categorized by **F0 range**
---
## 🚀 Requirements
To use this model **without errors**, you **must** use **Applio v3.2.9 or later**, specifically versions **updated after July 15**.
Even if your Applio version is labeled 3.2.9, models updated **before** July 15 may still result in `mismatch` errors.
---
## 🌟 Key Features
- ✅ **595 F0-Categorized Speakers**
- ✅ **11,000+ hours of curated training data**
- ✅ **Spin Embedder** for flexible and generalizable voice transfer
- ✅ Real-time inference support
- ✅ Extensive support for **non-verbal sounds** such as:
- **Coughs**
- **Laughter**
- **Whispers**
- And other expressive human vocal behaviors
These features are made possible by including a large proportion of **non-speech data** in the training set.
---
## 📊 Recommended Usage
To fully utilize BlackTone’s capabilities—especially non-speech inference—you should include **10–20% non-speech data** in your fine-tuning dataset.
If you **do not** have such data, you can try setting the **Feature Index** to `0`, which may enable limited inference of non-verbal sounds. However, **non-speech training data is highly recommended** for best results.
---
## 📎 License
Apache 2.0
---
## 📫 Contact
For usage reports, contributions, or collaborations, please open an issue or contact the model maintainer.
|