Wav2vec2 Large Xlsr 53 Icelandic Ep30 967h

🚀 wav2vec2-large-xlsr-53-icelandic-ep30-967h

The "wav2vec2-large-xlsr-53-icelandic-ep30-967h" is an acoustic model designed for Automatic Speech Recognition in Icelandic. It's derived from fine - tuning the facebook/wav2vec2-large-xlsr-53 model over 30 epochs, using 967 hours of Icelandic data. This data was collected by the Language and Voice Laboratory via the Samrómur platform.

🚀 Quick Start

The "wav2vec2-large-xlsr-53-icelandic-ep30-967h" can be readily used for Icelandic Automatic Speech Recognition tasks. It's built upon fine - tuning a pre - trained model with a large amount of Icelandic data, ensuring high - quality performance.

✨ Features

• Icelandic - Specific: Tailored for Automatic Speech Recognition in Icelandic.
• Fine - Tuned: Based on facebook/wav2vec2-large-xlsr-53, fine - tuned with 967 hours of Icelandic data.
• Diverse Datasets: Trained on the Samrómur Milljón corpus.

📦 Installation

No explicit installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

import torch
from transformers import Wav2Vec2Processor
from transformers import Wav2Vec2ForCTC

#Load the processor and model.
MODEL_NAME="language-and-voice-lab/wav2vec2-large-xlsr-53-icelandic-ep30-967h"
processor = Wav2Vec2Processor.from_pretrained(MODEL_NAME)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_NAME)

#Load the dataset
from datasets import load_dataset, load_metric, Audio
ds=load_dataset("language-and-voice-lab/samromur_children", split="test")

#Downsample to 16kHz
ds = ds.cast_column("audio", Audio(sampling_rate=16_000))

#Process the dataset
def prepare_dataset(batch):
    audio = batch["audio"]
    #Batched output is "un-batched" to ensure mapping is correct
    batch["input_values"] = processor(audio["array"], sampling_rate=audio["sampling_rate"]).input_values[0]
    with processor.as_target_processor():
        batch["labels"] = processor(batch["normalized_text"]).input_ids
    return batch
ds = ds.map(prepare_dataset, remove_columns=ds.column_names,num_proc=1)

#Define the evaluation metric
import numpy as np
wer_metric = load_metric("wer")
def compute_metrics(pred):
    pred_logits = pred.predictions
    pred_ids = np.argmax(pred_logits, axis=-1)
    pred.label_ids[pred.label_ids == -100] = processor.tokenizer.pad_token_id
    pred_str = processor.batch_decode(pred_ids)
    #We do not want to group tokens when computing the metrics
    label_str = processor.batch_decode(pred.label_ids, group_tokens=False)
    wer = wer_metric.compute(predictions=pred_str, references=label_str)
    return {"wer": wer}

#Do the evaluation (with batch_size=1)
model = model.to(torch.device("cuda"))
def map_to_result(batch):
    with torch.no_grad():
        input_values = torch.tensor(batch["input_values"], device="cuda").unsqueeze(0)
        logits = model(input_values).logits
    pred_ids = torch.argmax(logits, dim=-1)
    batch["pred_str"] = processor.batch_decode(pred_ids)[0]
    batch["sentence"] = processor.decode(batch["labels"], group_tokens=False)
    return batch
results = ds.map(map_to_result,remove_columns=ds.column_names)

#Compute the overall WER now.
print("Test WER: {:.3f}".format(wer_metric.compute(predictions=results["pred_str"], references=results["sentence"])))

Test Result: 0.076

📚 Documentation

Model Details

The model "wav2vec2-large-xlsr-53-icelandic-ep30-967h" is fine - tuned from facebook/wav2vec2-large-xlsr-53. The fine - tuning was carried out for 30 epochs using 967 hours of Icelandic data from the Samrómur Milljón corpus.

Evaluation Results

📄 License

This model is released under the CC - BY - 4.0 license.

📖 BibTeX entry and citation info

When publishing results based on these models please refer to:

@inproceedings{mena2024samromur,
  title={Samr{\'o}mur Millj{\'o}n: An ASR Corpus of One Million Verified Read Prompts in Icelandic},
  author={Mena, Carlos Daniel Hernandez and Gunnarsson, {\TH}orsteinn Da{\dh}i and Gu{\dh}nason, J{\'o}n},
  booktitle={Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)},
  pages={14305--14312},
  year={2024}
}

🙏 Acknowledgements

• Thanks to Jón Guðnason, head of the Language and Voice Lab, for providing computational power.
• Thanks to the "Language Technology Programme for Icelandic 2019 - 2023", managed and coordinated by Almannarómur and funded by the Icelandic Ministry of Education, Science and Culture.
• Special thanks to Björn Ingi Stefánsson for setting up the server configuration for model training.