Wav2vec2 Large Tedlium

🚀 Wav2Vec2-Large-Tedlium

A fine-tuned Wav2Vec2 large model on the TEDLIUM corpus for speech recognition.

This model is initialized with Facebook's Wav2Vec2 large LV-60k checkpoint, which is pre-trained on 60,000 hours of audiobooks from the LibriVox project. It is fine-tuned on 452 hours of TED talks from the TEDLIUM corpus (Release 3). When using the model, ensure that your speech input is sampled at 16Khz.

The model achieves a word error rate (WER) of 8.4% on the dev set and 8.2% on the test set. The Training logs document the training and evaluation progress over 50k steps of fine-tuning.

For more information on how this model was fine-tuned, see this notebook.

🚀 Quick Start

Prerequisites

• Ensure your speech input is sampled at 16Khz.

Transcribing Audio Files

The model can be used as a standalone acoustic model to transcribe audio files.

💻 Usage Examples

Basic Usage

from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
from datasets import load_dataset
import torch

# load model and processor
processor = Wav2Vec2Processor.from_pretrained("sanchit-gandhi/wav2vec2-large-tedlium")
model = Wav2Vec2ForCTC.from_pretrained("sanchit-gandhi/wav2vec2-large-tedlium")

# load dummy dataset
ds = load_dataset("sanchit-gandhi/tedlium_dummy", split="validation")

# process audio inputs
input_values = processor(ds[0]["audio"]["array"], return_tensors="pt", padding="longest").input_values  # Batch size 1

# retrieve logits
logits = model(input_values).logits

# take argmax and decode
predicted_ids = torch.argmax(logits, dim=-1)
transcription = processor.batch_decode(predicted_ids)
print("Target: ", ds["text"][0])
print("Transcription: ", transcription[0])

Evaluation

The following code snippet shows how to evaluate Wav2Vec2-Large-Tedlium on the TEDLIUM test data.

Advanced Usage

from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import torch
from jiwer import wer

tedlium_eval = load_dataset("LIUM/tedlium", "release3", split="test")
model = Wav2Vec2ForCTC.from_pretrained("sanchit-gandhi/wav2vec2-large-tedlium").to("cuda")
processor = Wav2Vec2Processor.from_pretrained("sanchit-gandhi/wav2vec2-large-tedlium")
def map_to_pred(batch):
    input_values = processor(batch["audio"]["array"], return_tensors="pt", padding="longest").input_values
    with torch.no_grad():
        logits = model(input_values.to("cuda")).logits
    predicted_ids = torch.argmax(logits, dim=-1)
    transcription = processor.batch_decode(predicted_ids)
    batch["transcription"] = transcription
    return batch
result = tedlium_eval.map(map_to_pred, batched=True, batch_size=1, remove_columns=["speech"])
print("WER:", wer(result["text"], result["transcription"]))

📄 License

This project is licensed under the Apache-2.0 license.