Fine-tuned F5-TTS Model on tunisian dialect

    This is a fine-tuned checkpoint of [SWivid/F5-TTS](https://huggingface.co/SWivid/F5-TTS),
    a flow-matching based text-to-speech model.

    ## Files

    - `model_last.pt` — fine-tuned model checkpoint
    - `vocab.txt` — tokenizer vocabulary

• F5TTS_v1_Base_TUN.yaml — training/model config

    ## Usage
  
    Load this checkpoint with the F5-TTS inference code
    (https://github.com/SWivid/F5-TTS):
  
    ```bash
    f5-tts_infer-cli \
        --model F5-TTS \
        --ckpt_file model_last.pt \
        --vocab_file vocab.txt \
        --ref_audio "path/to/reference.wav" \
        --ref_text "Transcript of the reference audio." \
        --gen_text "Text you want the model to speak."
    ```
  
    Or in Python:
  
    ```python
    from f5_tts.api import F5TTS
  
    tts = F5TTS(
        model="F5TTS_v1_Base",
        ckpt_file="model_last.pt",
        vocab_file="vocab.txt",
    )
    wav, sr, spect = tts.infer(
        ref_file="path/to/reference.wav",
        ref_text="Transcript of the reference audio.",
        gen_text="Text you want the model to speak.",
    )
    ```
  
    ## Training details
  
    _Fill in: dataset used, number of steps/epochs, hardware, base checkpoint, etc._
  
    ## License
  
    Inherits the license of the base model unless noted otherwise above.