Name: Tyla | AI RVC Model
Brand: Weights
Rating: 4.0 (34 reviews)

WebKit

 Introducing Tyla (RVC v2) 500 Epochs, the advanced AI voice model by Weights that uses Retrieval-Based Voice Conversion (RVC) technology to provide realistic and high-quality text-to-speech and music generation capabilities. This innovative model has been trained using the cutting-edge RMVPE extraction method for 500 epochs, resulting in enhanced precision and accuracy in its outputs. With a dataset length of 04:12 extracted from popular songs, Tyla (RVC v2) 500 Epochs offers remarkable versatility, allowing you to generate unique AI covers of your favorite tunes effortlessly. Try out our free AI tools today and experience the power of Tyla's exceptional abilities for yourself!

Creator / Dataset Extraction method: RMVPE Epochs: 500 Dataset length: 04:12 (from available songs on TYLA + "Getting Late") : From testing the model I would recommend using RMVPE / RMVPE+ for inference. Raw acapella vocals work BEST for this model. All audios except for needy were inferred with raw vocals; however processed vocals still work. My Models Brett Cooper (RVC v2) 250 Epochs Jacob Zuma (RVC v2) 200 Epochs

Tyla

spiiidey

- **Creator / Dataset Creator**: <@966717388169052160> 

- **Extraction method**: RMVPE 

- **Epochs**: 500

- **Dataset length**: 04:12 (from available songs on “TYLA” + "Getting Late")

- **Link**: https://huggingface.co/spideyyy/ZA_Voices/blob/main/Tyla.zip

- **Note**: From testing the model, I would recommend using RMVPE / RMVPE+ for inference. 

Raw acapella vocals work BEST for this model. All audios except for “needy” were inferred with raw vocals; however, processed vocals still work. 

------------------------------------------------------------------------------

**__My Models__**
- Brett Cooper (RVC v2) 250 Epochs
- Jacob Zuma (RVC v2) 200 Epochs

- Creator / Dataset - Extraction method: RMVPE - Epochs: 500 - Dataset length: 06:33 (from available songs on P2: The Big Hearted Bad Guy + others) - Link Note: From testing the model, I would recommend using RMVPE / RMVPE+ for inference [I've found that RMVPE+ works best] . Raw acapella vocals work BEST for this model. DM me for inquiries about private models __My Models__ - Brett Cooper (RVC v2) 250 Epochs - Jacob Zuma (RVC v2) 200 Epochs -

A-Reece

A-Reece (RVC v2) 500 Epochs

Training mode: Harvest Dataset length: 5 minutes 38 seconds Taken from: Link Dataset collected and model trained by me. - One TTS demo. - One speech-to-speech demo. - One singing demo.

Brett Cooper

Brett Cooper (RVC v2) 250 Epochs

collected and trained by me (via harvest). Dataset taken from parliamentary speeches and then enhanced with Adobe Podcast + trained with 10:06 minutes of audio (mostly English audio but also some isiZulu audio; however, clicks DO NOT work as far as I'm concerned - I could update that if necessary). Only trained on speaking (no singing; however, I will send a singing demo below)

Tyla

Description

Samples

Tyla (South African Singer)

Tyla

Tyla (Tyla +) (2024)

Tyla

Dach Barry (DACHTYLA STRAVROMENA)