license: cc-by-nc-4.0
datasets:
- mozilla-foundation/common_voice_17_0
- facebook/voxpopuli
- mrfakename/librivox-full-catalog-archive
language:
- fi
base_model:
- SWivid/F5-TTS
pipeline_tag: text-to-speech
Here are three Finnish models of the F5-TTS, listen speech samples for models.
Numbers cannot be understood by models. Convert numbers to words.
The Common Voice and Vox Populi Finnish datasets are used for the first round.
20241206
Speakers: Several speakers from different corpus
Use these with "f5-tts_infer-gradio":
Model: hf://AsmoKoskinen/F5-TTS_Finnish_Model/model_common_voice_fi_vox_populi_fi_20241206.safetensors
Vocab: hf://AsmoKoskinen/F5-TTS_Finnish_Model/vocab.txt
The second round is based on the Common Voice, LibriVox and Vox Populi Finnish data sets. Use this as a default one.
20241217
Speakers: Several speakers from different corpus
Use these with "f5-tts_infer-gradio":
Model: hf://AsmoKoskinen/F5-TTS_Finnish_Model/model_commonvoice_fi_librivox_fi_vox_populi_fi_20241217/model_last_20241217.safetensors
Vocab: hf://AsmoKoskinen/F5-TTS_Finnish_Model/model_commonvoice_fi_librivox_fi_vox_populi_fi_20241217/vocab.txt
The third round is based on the Common Voice, LibriVox and Vox Populi Finnish data sets, same as the second round. This one is no better.
20250125
Speakers: Several speakers from different corpus
Use these with "f5-tts_infer-gradio":
Model: hf://AsmoKoskinen/F5-TTS_Finnish_Model/model_commonvoice_fi_librivox_fi_vox_populi_fi_20250125/model_last_20250125.safetensors
Vocab: hf://AsmoKoskinen/F5-TTS_Finnish_Model/model_commonvoice_fi_librivox_fi_vox_populi_fi_20250125/vocab.txt