whisper-large-v3 Model by OpenAI

Getting Started

Riva uses gRPC APIs. Instructions below demonstrate usage of whisper-large-v3 model using Python gRPC client.

Prerequisites

You will need a system with Git and Python 3+ installed.

Install Riva Python Client

pip install nvidia-riva-client

Download Python Client

Download Python client code by cloning Python Client Repository.

git clone https://github.com/nvidia-riva/python-clients.git

Run Python Client

Make sure you have a speech file in Mono, 16-bit audio in WAV, OPUS and FLAC formats. If you have generated the API key, it will be auto-populated in the command. Open a command terminal and execute below command to transcribe audio. Specifying --language-code as multi will enable auto language detection. If you know the source language, it is recommended to specify for better accuracy and latency. See Supported Languages for the list of all available languages and corresponding code.

python python-clients/scripts/asr/transcribe_file_offline.py \
    --server grpc.nvcf.nvidia.com:443 --use-ssl \
    --metadata function-id "b702f636-f60c-4a3d-a6f4-f3568c13bd7d" \
    --metadata "authorization" "Bearer $NVIDIA_API_KEY" \
    --language-code en \
    --input-file <path_to_audio_file>

Below command demonstrates translation from French (fr) to English.

python python-clients/scripts/asr/transcribe_file_offline.py \
    --server grpc.nvcf.nvidia.com:443 --use-ssl \
    --metadata function-id "b702f636-f60c-4a3d-a6f4-f3568c13bd7d" \
    --metadata "authorization" "Bearer $NVIDIA_API_KEY" \
    --language-code fr \
    --custom-configuration "task:translate" \
    --input-file <path_to_audio_file>

Support for gRPC clients in other languages

Riva uses gRPC APIs. Proto files can be downloaded from Riva gRPC Proto files and compiled to target language using Protoc compiler. Example Riva clients in C++ and Python languages are provided below.

openai

whisper-large-v3