You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this model content.

⚠️ WORK IN PROGRESS: This model is still in early training stages. Current checkpoints produce low-quality, barely distinguishable speech. Updated checkpoints will be released as training progresses.

OpenF5-TTS

A commercial-friendly version of F5-TTS retrained from scratch on permissively-licensed data.

Trained on the Emilia-YODAS (CC-BY) dataset using the F5-TTS Small configuration.

GitHub Repository + Details

Usage

The model requires specific configuration files to work properly:

pip install f5-tts
huggingface-cli download mrfakename/OpenF5-Intermediate --local-dir openf5
f5-tts_infer-cli -mc openf5/model_config.yaml -p openf5/model_last.pt -v openf5/vocab.txt

Training Progress

Listen to various audio samples across steps. All samples use this text:

I don't really care what you call me. I've been a silent spectator, watching species evolve, empires rise and fall. But always remember, I am mighty and enduring.

Reference audio:

~250K steps (Current Checkpoint)

~200K steps

~150K steps

(Starting to hear traces of the expected text!)

~100K steps

~75K steps

~50K steps

~25K steps

Support

Discussions are disabled until training is complete. For issues, please open a GitHub Issue in the repository.

License

  • Model: CC-BY 4.0 - Free for commercial use
  • Scripts: MIT License

Note: No restrictions are placed on usage of the outputs of the model. While attribution is appreciated, it is not required for outputs of the model.

THE MODEL IS PROVIDED “AS IS” UNDER ITS OPEN-SOURCE LICENSE. THE AUTHORS AND CONTRIBUTORS DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT. USERS ARE SOLELY RESPONSIBLE FOR ENSURING COMPLIANCE WITH APPLICABLE COPYRIGHT LAWS, INCLUDING THE USE OF INPUT DATA AND GENERATED OUTPUTS. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES, INCLUDING BUT NOT LIMITED TO DAMAGES RESULTING FROM LOSS OF USE, DATA, OR PROFITS, OR ANY CLAIMS RELATED TO THE MODEL’S OUTPUTS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE, OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.