ThaiVanPhat95/wav2vec2-large-l2-arctic-supcon-repeated-8fold-0-4gram

Wav2Vec2 Large L2-ARCTIC SupCon Repeated: 8-Fold Split 0 with 4-Gram LM

Associated paper: Contrastive Regularization for Accent-Robust ASR.

English accented-speech recognition model trained with CTC and supervised contrastive learning. Repeated L2-ARCTIC prompts provide positive SupCon pairs. This package adds the available Arctic 4-gram language model for decoding.

Data and Code

• —L2-ARCTIC: https://psi.engr.tamu.edu/l2-arctic-corpus/
• —Training code and split metadata: https://github.com/thaivanphat95/robust-atc-asr

Usage

pip install pyctcdecode

import torch
import soundfile as sf
from transformers import AutoModelForCTC, Wav2Vec2ProcessorWithLM

model_id = "thaivanphat95/wav2vec2-large-l2-arctic-supcon-repeated-8fold-0-4gram"
processor = Wav2Vec2ProcessorWithLM.from_pretrained(model_id)
model = AutoModelForCTC.from_pretrained(model_id).eval()

audio, sample_rate = sf.read("audio.wav")
inputs = processor(audio, sampling_rate=sample_rate, return_tensors="pt")

with torch.no_grad():
    logits = model(input_values=inputs.input_values).logits

print(processor.decode(logits.cpu().numpy()[0]).text)

Audio should be mono and sampled at 16 kHz.

Citation

@article{thai2026contrastive,
  title={Contrastive Regularization for Accent-Robust ASR},
  author={Thai, Van-Phat and Dhruv, Aradhya and Pham, Duc-Thinh and Alam, Sameer},
  journal={arXiv preprint arXiv:2605.03297},
  year={2026},
  doi={10.48550/arXiv.2605.03297}
}

Limitations and License

The acoustic and language models are specialized for L2-ARCTIC-style English read speech. Language-model training data should be considered when interpreting split-specific evaluation. Model-weight and LM use may be affected by the licenses of the pretrained model and L2-ARCTIC. The training code's Apache-2.0 license does not cover these artifacts.