SpireLab/Syspin_Hackathon_api_server

SYSPIN Hackathon TTS API Documentation

Overview

This API provides a Text-to-Speech (TTS) service that converts input text into speech audio. It supports multiple Indian languages and offers voice customization through speaker references provided by the user .

---

Endpoint: /Get_Inference

• Method: GET
• Description: Generates speech audio from the provided text using the specified language and speaker reference file.

Query Parameters

Parameter	Type	Required	Description
text	string	Yes	The input text to be converted into speech.
lang	string	Yes	The language of the input text. Acceptable values include: bhojpuri, bengali, english, gujarati, hindi, chhattisgarhi, kannada, magahi, maithili, marathi, telugu.
speaker_wav	WAV file (Bytes)	Yes	Must be a WAV file

Available Languages

Language	Language codes
chhattisgarhi	hne
kannada	kn
maithili	mai
telugu	te
bengali	bn
bhojpuri	bho
marathi	mr
gujarati	gu
hindi	hi
magahi	mag
english	en

Responses

• 200 OK: Returns a WAV audio file as a streaming response containing the synthesized speech.

• 422 Unprocessable Entity: Returned when:

  • Any of the required query parameters (text, lang, speaker_wav) are missing.
  • The specified lang is not supported.
  • The specified speaker_wav is not available.

Running the Server

To start the FastAPI server:

docker build -t your_image_name ./
docker run -d -v /path/to/this/code/dir/:/app/ -p 8080:8080 your_image_name API_main.py

Hosting on a GPU

To run your FastAPI-based Text-to-Speech (TTS) server inside a Docker container with GPU support, follow these steps:

---

Prerequisites

1. NVIDIA GPU: Ensure your system has an NVIDIA GPU installed.

2. NVIDIA Drivers: Install the appropriate NVIDIA drivers for your GPU.

3. Docker: Install Docker on your system.

4. NVIDIA Container Toolkit: Install the NVIDIA Container Toolkit to enable GPU support in Docker containers.

---

Installation Steps

1. Install NVIDIA Drivers

Ensure that the NVIDIA drivers compatible with your GPU are installed on your system.

2. Install Docker

If Docker is not already installed, you can install it by following the official Docker installation guide for your operating system.

3. Install NVIDIA Container Toolkit

The NVIDIA Container Toolkit allows Docker containers to utilize the GPU.

For Ubuntu:

# Add the package repositories
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | \
  sudo tee /etc/apt/sources.list.d/nvidia-docker.list

# Update the package lists
sudo apt-get update

# Install the NVIDIA Container Toolkit
sudo apt-get install -y nvidia-container-toolkit

# Restart the Docker daemon to apply changes
sudo systemctl restart docker

For other operating systems: Refer to the NVIDIA Container Toolkit installation guide for detailed instructions.

4. Verify GPU Access in Docker

To confirm that Docker can access your GPU, run the following command:

docker run --rm --gpus all nvidia/cuda:12.4.0-base-ubuntu22.04 nvidia-smi

Running Your FastAPI TTS Server with GPU Support

Assuming your FastAPI TTS application is containerized and ready to run:

1. Build Your Docker Image

Navigate to the directory containing your Dockerfile and build the Docker image:

docker build -t your_image_name .

2. Run the Docker Container with GPU Support

Start the container with GPU access enabled:

docker run --gpus all -p 8080:8080 -v /path/to/this/code/dir/:/app/ your_image_name API_main.py

Example API Call

import requests

# Define the base URL of your API
base_url = 'http://localhost:8080/Get_Inference'

# Set up the query parameters
WavPath = 'path/to/wavfile.wav'

params = {
    'text': 'ಮಾದರಿಯು ಸರಿಯಾಗಿ ಕಾರ್ಯನಿರ್ವಹಿಸುತ್ತಿದೆಯೇ ಎಂದು ಖಚಿತಪಡಿಸಿಕೊಳ್ಳಲು ಬಳಸಲಾಗುವ ಪರೀಕ್ಷಾ ವಾಕ್ಯ ಇದು.',
    'lang': 'kannada',
}

# Send the GET request
with open(WavPath, "rb") as AudioFile:
  response = requests.get(base_url, params = params, files = { 'speaker_wav': AudioFile.read() })

# Check if the request was successful
if response.status_code == 200:
    # Save the audio content to a file
    with open('output.wav', 'wb') as f:
        f.write(response.content)
    print("Audio saved as 'output.wav'")
else:
    # Print the error message
    print(f"Request failed with status code {response.status_code}")
    print("Response:", response.text)