Skip to content

Commit

Permalink
update readme with how to run code
Browse files Browse the repository at this point in the history
  • Loading branch information
owos committed Mar 4, 2024
1 parent b5d30e8 commit 581cf50
Showing 1 changed file with 53 additions and 0 deletions.
53 changes: 53 additions & 0 deletions recipes/ljspeech/xtts_v2/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,53 @@

# Coqui AI TTS

This repository contains the code for training a text-to-speech (TTS) model using Coqui AI's TTS framework.

## Installation and Configuration

1. **Clone the repository:**

```bash
git clone https://github.com/owos/coqui-ai-TTS.git
```

2. **Navigate to the repository root:**

```bash
cd coqui-ai-TTS
```

3. **Install system dependencies and the code:**

```bash
make system-deps # Intended to be used on Ubuntu (Debian). Let us know if you have a different OS.
make install
```

4. **Open the following file and redefine the specified variables:**

File: `recipes/ljspeech/xtts_v2/train_gpt_xtts.py`

```python
# Line 30
path = 'the root path to the audio dirs on your machine'

# Line 31
meta_file_train = "the root path to the train CSV on your machine"

# Line 32
meta_file_val = "the root path to the train CSV on your machine"

# Line 75
SPEAKER_REFERENCE = "a list with a single path to a test audio from the afro tts data"
```

## Running the Code

From the repository root, run the following command:

```bash python
python3 recipes/ljspeech/xtts_v2/train_gpt_xtts.py
```

You are now ready to train your TTS model using Coqui AI's framework. Enjoy!

0 comments on commit 581cf50

Please sign in to comment.