Skip to content
AI Tutorials
Text-to-Speech
F5-TTS

Install SWivid/F5-TTS with Conda

Introduction

F5-TTS is a flow matching-based text-to-speech synthesis system that generates high-quality speech from text input. The system implements Conditional Flow Matching (CFM) with transformer-based architectures to produce natural-sounding speech with voice cloning capabilities.

Model Architecture Backbone Description
F5-TTS Diffusion Transformer DiT, MMDiT ConvNeXt V2-based, optimized for speed
E2-TTS Flat-UNet Transformer UNetT Faithful reproduction of E2-TTS paper

GitHub: https://github.com/swivid/f5-tts
Hugging Face: https://hf.co/SWivid/F5-TTS
Paper: https://hf.co/papers/2410.06885
Demo: https://hf.co/spaces/mrfakename/E2-F5-TTS

Prerequisites

System requirements:

  • Operating System: Windows 10/11 (64-bit), macOS, or Linux (Debian/Ubuntu).
  • Python: Version >=3.10 required
  • Disk Space: 6GB+ recommended (for dependencies and model cache). At least 400 MB for Miniconda; 3 GB+ for full Anaconda.
  • The GPU is optional but HIGHLY Recommended for Performance
  • Internet: For downloading dependencies and models from Hugging Face Hub.
Requirements CPU RAM
Minimum Intel Core i3 or AMD Ryzen 3 8 GB
Recommended Intel Core i5 or AMD Ryzen 5 16 GB
High-performance Intel Core i7/i9 or AMD Ryzen 7/9 32 GB
Environment Run this Command
CPU only pip3 install torch torchvision
Apple Silicon pip3 install torch torchvision
AMD ROCm Install PyTorch for ROCm
CUDA 11.8 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
CUDA 12.1 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
CUDA 12.6 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126
CUDA 12.8 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
CUDA 13.0 pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu130

Video tutorial

Step 1. Install Miniconda Package

Step 2. Create Conda Environment

Create a conda environment:

name: f5-tts
channels:
- conda-forge
- defaults
dependencies:
# Python version >= 3.10 required
- python=3.11

# Install FFmpeg if you haven't yet
- ffmpeg

- pip
- pip:
    # PyTorch CPU only
    - torch
    - torchaudio

    # Allow run Gradio app (web interface)
    - gradio

    # F5-TTS
    - f5-tts

Activate conda environment:

conda env create -f environment.yml
conda activate f5-tts

Note: RuntimeWarning: Couldn't find ffmpeg or avconv - defaulting to ffmpeg, but may not work can be safely ignored if you only work with WAV audio files, as Pydub supports reading and writing WAV files natively without FFmpeg. However, FFmpeg is required for processing other audio formats such as MP3, OGG, and FLAC, as well as for audio format conversion.

Step 3. Run the Inference

CLI Inference

Reference audio sample voice:

# Run with flags
# Leave --ref_text "" will have ASR model transcribe (extra GPU memory usage)
f5-tts_infer-cli --model F5TTS_v1_Base \
--ref_audio "D:\f5-tts\ref_audio.wav" \
--ref_text "The piper was very glad to see the pig and said to Tom" \
--gen_text "Some text you want TTS model generate for you."

The result will be the audio file tests\infer_cli_basic.wav

Gradio app (web interface) 

# Launch a Gradio app (web interface)
f5-tts_infer-gradio

Starting app…

  • Running on local URL: http://127.0.0.1:7860

F5-TTS

dots.ttsICE-012