You can deploy this model and start earning money today!

mahesh27/
w2v2l-custom-archi

Name: w2v2l-custom-archi
Availability: InStock

safetensors

wav2vec2

apache-2.0

Views

No views yet

Automatic Speech Recognition model for Archi, IPA output

Example Usage

python

from datasets import load_dataset
from transformers import AutoProcessor, AutoModelForCTC
import torch

# Streaming dataset
ds = load_dataset(
    "mahesh27/archi_rutul_asr",
    "Archi",
    split="train",
    streaming=True,
)

# First example
sample = next(iter(ds))

processor = AutoProcessor.from_pretrained("mahesh27/w2v2l-custom-archi")
model = AutoModelForCTC.from_pretrained(
    "mahesh27/w2v2l-custom-archi",
    device_map="auto",
)

# Prepare input
inputs = processor(
    sample["audio"]["array"],
    sampling_rate=sample["audio"]["sampling_rate"],
    return_tensors="pt",
    padding=True,
)

# Move tensors to model device
inputs = {k: v.to(model.device) for k, v in inputs.items()}

model.eval()
with torch.no_grad():
    logits = model(**inputs).logits

pred_ids = torch.argmax(logits, dim=-1)

transcription = processor.batch_decode(pred_ids)[0]

print("Reference :", sample["transcript"])
print("Prediction:", transcription)

Reference : qʼˤʷebuʃi abutːib lobur
Prediction: eˤbaʃi abitːib lobur

Citation:

@inproceedings{akavarapu-etal-2026-hard,
    title = "Hard to Be Heard: Phoneme-Level {ASR} Analysis of Phonologically Complex, Low-Resource Endangered Languages",
    author = {Akavarapu, V.S.D.S.Mahesh  and
      Daniel, Michael  and
      J{\"a}ger, Gerhard},
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.findings-acl.147/",
    doi = "10.18653/v1/2026.findings-acl.147",
    pages = "3014--3028",
    ISBN = "979-8-89176-395-1",
    abstract = "We present a phoneme-level analysis of automatic speech recognition (ASR) for two low-resourced and phonologically complex East Caucasian languages, Archi and Rutul, based on curated and standardized speech{--}transcript resources totaling approximately 50 minutes and 1 hour 20 minutes of audio, respectively. Existing recordings and transcriptions are consolidated and processed into a form suitable for ASR training and evaluation. We evaluate several state-of-the-art audio and audio{--}language models, including wav2vec2, Whisper, and Qwen2-Audio. For wav2vec2, we introduce a language-specific phoneme vocabulary with heuristic output-layer initialization, which yields consistent improvements and achieves performance comparable to or exceeding Whisper in these extremely low-resource settings. Beyond standard word and character error rates, we conduct a detailed phoneme-level error analysis. We find that phoneme recognition accuracy strongly correlates with training frequency, exhibiting a characteristic sigmoid-shaped learning curve. For Archi, this relationship partially breaks for Whisper, pointing to model-specific generalization effects beyond what is predicted by training frequency. Overall, our results indicate that many errors attributed to phonological complexity are better explained by data scarcity. These findings demonstrate the value of phoneme-level evaluation for understanding ASR behavior in low-resource, typologically complex languages."
}