tonythethompson/OpenF5-TTS-Base

OpenF5 TTS Base (Alpha) — mirror

Source

Field	Value
Upstream model	mrfakename/OpenF5-TTS-Base
Upstream source revision	28e45df3acf0595c9febf233d7d82b409708ce2c
Export tool/script	Weight mirror from upstream PyTorch checkpoint (no conversion)
Quantization recipe	FP32 PyTorch (model.pt)

Precision and Packaging

Export tooling, precision, and quantization are recorded in the Source table above. This packaging mirror does not publish independent parity benchmarks; validate on your target execution provider before production use.

Upstream details (per mrfakename's card)

• Trained (by mrfakename) using the F5-TTS Base V1 configuration for ~1 million steps on the Emilia-YODAS dataset. English speech only.

Usage

This is a mirror; you can pull from either repo. Example (upstream):

pip install f5-tts
huggingface-cli download mrfakename/OpenF5-TTS-Base --local-dir openf5
f5-tts_infer-cli -mc openf5/config.yaml -p openf5/model.pt -v openf5/vocab.txt

Replace the repo id with tonythethompson/OpenF5-TTS-Base to pull from this mirror instead.

Safety and Responsible Use

This model can perform zero-shot voice cloning and produce realistic synthetic speech.

• Do not use to impersonate real individuals without their explicit consent.
• Do not generate synthetic speech intended to deceive listeners about a speaker's identity.
• Disclose AI-generated audio where listeners would reasonably expect a human voice.
• Users are responsible for compliance with applicable laws governing synthetic media and voice cloning in their jurisdiction.

Acknowledgements

Model and training by mrfakename. Thanks to the authors of F5-TTS, to lucidrains for the original E2-TTS implementation, and to Amphion for the Emilia-YODAS dataset.

License

Apache 2.0 — same as the upstream mrfakename/OpenF5-TTS-Base. Usable for commercial and non-commercial purposes. This mirror adds no new license terms.