SPEAR XLarge (speech + general audio)
Recommended: This is the original SPEAR XLarge v1 checkpoint. For new projects, we recommend using the ICML 2026 accepted XLarge v2 model: marcoyang/spear-xlarge-speech-audio-v2. XLarge v2 is aligned with the accepted ICML 2026 version and includes additional robustness improvements for complex acoustic scenes through token mixing.
This is the SPEAR XLarge v1 dual-domain (speech + general audio) model. The model adopts a Zipformer backbone with 597M parameters consisting of 13 Zipformer stacks. It generates 1280-dimensional representations at approximately 50~Hz.
The model was pre-trained on 197k hours of mixture data of English speech and general audio, among which 184k hours are speech data, and the rest 13k hours are general audio data. It achieves state-of-the-art performance on the SUPERB benchmark and competitive performance on HEAR. For the latest SPEAR XLarge model with stronger complex-scene robustness, please use XLarge v2.
The speech data consists of the following datasets:
| Dataset | Duration (hours) |
|---|---|
| Libriheavy | ~50k |
| Gigaspeech | ~10k |
| VoxPopuli (en) | ~24k |
| yodas-granary | ~100k |
The audio data consists of the following datasets:
| Dataset | Duration (hours) |
|---|---|
| AudioSet | ~5k |
| Freesound | ~2.8k |
| Music4all | ~1k |
| VGGSound | ~ 0.5k |
| MTG-Jamendo | ~ 3.8k |
Note: The model is pretrained on 16kHz sampled speech/audio data. When using the model make sure that your input is also sampled at 16kHz.
Authors: Xiaoyu Yang, Yifan Yang, Zengrui Jin, Ziyun Cui, Wen Wu, Baoxiang Li, Chao Zhang, Phil Woodland
Abstract Self-Supervised Learning (SSL) excels at learning generic representations of acoustic signals, yet prevailing methods remain domain-specific, tailored to either speech or general audio, hindering the development of a unified representation model with a comprehensive capability over both domains. To address this, we present SPEAR (SPEech and Audio Representations), the first SSL framework to successfully learn unified speech and audio representations from a mixture of speech and audio data. SPEAR proposes a unified pre-training objective based on masked prediction of fine-grained discrete tokens for both speech and general audio. These tokens are derived from continuous speech and audio representations using a Multi-codebook Vector Quantisation (MVQ) method, retaining rich acoustic detail essential for modelling both speech and complex audio events. SPEAR is applied to pre-train both single-domain and unified speech-and-audio SSL models. Our speech-domain model establishes a new state-of-the-art on the SUPERB benchmark, a speech processing benchmark for SSL models, matching or surpassing the highly competitive WavLM Large on 12 out of 15 tasks with the same pre-training corpora and a similar model size. Crucially, our unified model learns complementary features and demonstrates comprehensive capabilities across two major benchmarks, SUPERB and HEAR, for evaluating audio representations. By further scaling up the model size and pre-training data, we present a unified model with 600M parameters that excels in both domains, establishing it as one of the most powerful and versatile open-source SSL models for auditory understanding.
Usage
This model is pre-trained purely using unlabelled data. Therefore, it requires fine-tuning with labelled data for downstream tasks such as automatic speech recognition (ASR) or audio tagging (AT).
The model achieves the following word error rates (WERs) when fine-tuned on LibriSpeech for ASR:
| Fine-tuning data | test-clean | test-other |
|---|---|---|
| LS960 | 1.6 | 2.9 |
The model acheives the following mean average precision (mAP) when fine-tuned on AudioSet for AT:
| Fine-tuning data | mAP |
|---|---|
| AudioSet Balanced | 39.4 |
| AudioSet Full | 50.0 |
You can extract its top-layer feature (and intermediate hidden states) using the following code:
from transformers import AutoModel
import torch
model = AutoModel.from_pretrained(
"marcoyang/spear-xlarge-speech-audio",
trust_remote_code=True,
force_download=False,
)
if torch.cuda.is_available():
model = model.to("cuda")
model.eval()
device = next(model.parameters()).device
audio = torch.randn(1, 160000).to(device) # dummy audio input of 10 seconds
audio_len = torch.tensor([160000]).to(device)
with torch.no_grad():
outputs = model(audio, audio_len)
encoder_out = outputs["encoder_out"] # (N,T,C)
encoder_out_lens = outputs["encoder_out_lens"] # (N)
middle_out = outputs["hidden_states"] # list of (N,T,C)
print(encoder_out)
print(encoder_out_lens)
print(len(middle_out)) # 13 layers
print(middle_out[-1].shape)
print(middle_out[-1])