Sew D Base Plus 400k Ft Ls100h

🚀 SEW-D-base+

A pre-trained speech model with high performance and efficiency for various speech tasks.

SEW-D-base+ is a base model pretrained on 16kHz sampled speech audio. When using the model, ensure that your speech input is also sampled at 16Khz. Note that this model should be fine - tuned on a downstream task, such as Automatic Speech Recognition, Speaker Identification, Intent Classification, Emotion Recognition, etc.

The original SEW - D model is developed by ASAPP Research and can be found at SEW - D by ASAPP Research.

Paper: Performance - Efficiency Trade - offs in Unsupervised Pre - training for Speech Recognition

Authors: Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi

Abstract This paper studies the performance - efficiency trade - offs in pre - trained models for automatic speech recognition (ASR). It focuses on wav2vec 2.0 and formalizes several architecture designs that influence both the model performance and its efficiency. By integrating all the observations, it introduces SEW (Squeezed and Efficient Wav2vec), a pre - trained model architecture with significant improvements in both performance and efficiency across various training setups. For example, under the 100h - 960h semi - supervised setup on LibriSpeech, SEW achieves a 1.9x inference speedup compared to wav2vec 2.0, with a 13.5% relative reduction in word error rate. With a similar inference time, SEW reduces the word error rate by 25 - 50% across different model sizes.

The original model checkpoints can be found at https://github.com/asappresearch/sew#model - checkpoints.

🚀 Quick Start

To transcribe audio files, the model can be used as a standalone acoustic model.

💻 Usage Examples

Basic Usage

from transformers import Wav2Vec2Processor, SEWDForCTC
from datasets import load_dataset
import soundfile as sf
import torch
 
# load the model and preprocessor
processor = Wav2Vec2Processor.from_pretrained("asapp/sew-d-base-plus-400k-ft-ls100h")
model = SEWDForCTC.from_pretrained("asapp/sew-d-base-plus-400k-ft-ls100h")

# load the dummy dataset with speech samples
ds = load_dataset("patrickvonplaten/librispeech_asr_dummy", "clean", split="validation")
 
# preprocess
input_values = processor(ds[0]["audio"]["array"], return_tensors="pt").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)

Advanced Usage

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

librispeech_eval = load_dataset("librispeech_asr", "clean", split="test")

model = SEWDForCTC.from_pretrained("asapp/sew-d-base-plus-400k-ft-ls100h").to("cuda")
processor = Wav2Vec2Processor.from_pretrained("asapp/sew-d-base-plus-400k-ft-ls100h")

def map_to_pred(batch):
    input_values = processor(batch["audio"][0]["array"], sampling_rate=16000, 
                             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 = librispeech_eval.map(map_to_pred, batched=True, batch_size=1, remove_columns=["audio"])

print("WER:", wer(result["text"], result["transcription"]))

📚 Documentation

Evaluation Results

Model Index

• Name: sew - d - base - plus - 400k - ft - ls100h
  • Results:
    • Task: Automatic Speech Recognition
      • Dataset: LibriSpeech (clean)
        • Metrics: Test WER = 4.34
    • Task: Automatic Speech Recognition
      • Dataset: LibriSpeech (other)
        • Metrics: Test WER = 9.45

📄 License

This project is licensed under the Apache - 2.0 license.