Wav2vec2 Base Da Ft Nst

🚀 wav2vec2-base-da-ft-nst

This is a Danish Automatic Speech Recognition (ASR) model based on alvenir wav2vec2 model, fine - tuned by Alvenir on the public NST dataset. It offers high - quality speech - to - text conversion for Danish.

🚀 Quick Start

This model is trained on 16kHz audio data. Ensure that your input data has the same sample rate. It was initially trained using fairseq and then converted to the huggingface/transformers format.

Alvenir is always willing to assist with your open - source ASR projects, customized domain specializations, or premium models. ;-)

✨ Features

• Danish ASR: Specifically fine - tuned for Danish language speech - to - text tasks.
• Sample Rate Requirement: Trained on 16kHz audio, ensuring compatibility with similar - sampled data.
• Format Compatibility: Converted to the huggingface/transformers format for easy integration.

📦 Installation

No specific installation steps are provided in the original README. If you want to use this model, you need to install relevant Python libraries such as transformers, soundfile, and torch. You can use the following command to install the transformers library:

pip install transformers

You may also need to install soundfile and torch according to your environment:

pip install soundfile torch

💻 Usage Examples

Basic Usage

import soundfile as sf
import torch

from transformers import Wav2Vec2CTCTokenizer, Wav2Vec2Tokenizer, Wav2Vec2Processor, \
    Wav2Vec2ForCTC


def get_tokenizer(model_path: str) -> Wav2Vec2CTCTokenizer:
    return Wav2Vec2Tokenizer.from_pretrained(model_path)


def get_processor(model_path: str) -> Wav2Vec2Processor:
    return Wav2Vec2Processor.from_pretrained(model_path)


def load_model(model_path: str) -> Wav2Vec2ForCTC:
    return Wav2Vec2ForCTC.from_pretrained(model_path)


model_id = "Alvenir/wav2vec2-base-da-ft-nst"

model = load_model(model_id)
model.eval()
tokenizer = get_tokenizer(model_id)
processor = get_processor(model_id)

audio_file = "<path/to/audio.wav>"

audio, _ = sf.read(audio_file)

input_values = processor(audio, return_tensors="pt", padding="longest", sampling_rate=16_000).input_values
with torch.no_grad():
    logits = model(input_values).logits

predicted_ids = torch.argmax(logits, dim=-1)
transcription = processor.batch_decode(predicted_ids)
print(transcription)

📚 Documentation

Benchmark results

Here are some benchmark results on publicly available Danish datasets.

📄 License

This project is licensed under the Apache - 2.0 license.