Competitive Audio-Language Models with Data-Efficient Single-Stage Training on Public Data

• URL: http://arxiv.org/abs/2509.07526v1
• Date: Tue, 09 Sep 2025 09:01:01 GMT
• Title: Competitive Audio-Language Models with Data-Efficient Single-Stage Training on Public Data
• Authors: Gokul Karthik Kumar, Rishabh Saraf, Ludovick Lepauloux, Abdul Muneer, Billel Mokeddem, Hakim Hacid,
• Abstract summary: Falcon3-Audio-7B matches the best reported performance among open-weight models on the MMAU benchmark.<n>Our smallest 1B model remains competitive with larger open models ranging from 2B to 13B parameters.
• Score: 4.736913024290765
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) have transformed NLP, yet their integration with audio remains underexplored -- despite audio's centrality to human communication. We introduce Falcon3-Audio, a family of Audio-Language Models (ALMs) built on instruction-tuned LLMs and Whisper encoders. Using a remarkably small amount of public audio data -- less than 30K hours (5K unique) -- Falcon3-Audio-7B matches the best reported performance among open-weight models on the MMAU benchmark, with a score of 64.14, matching R1-AQA, while distinguishing itself through superior data and parameter efficiency, single-stage training, and transparency. Notably, our smallest 1B model remains competitive with larger open models ranging from 2B to 13B parameters. Through extensive ablations, we find that common complexities -- such as curriculum learning, multiple audio encoders, and intricate cross-attention connectors -- are not required for strong performance, even compared to models trained on over 500K hours of data.

Related papers

• Scaling Open Discrete Audio Foundation Models with Interleaved Semantic, Acoustic, and Text Tokens [62.56027815951259]
  Current audio language models are predominantly text-first, either extending pre-trained text LLM backbones or relying on semantic-only audio tokens.<n>This paper presents a systematic empirical study of native audio foundation models that apply next-token prediction to audio at scale.
  arXiv  Detail & Related papers  (2026-02-18T18:32:46Z)
• Eureka-Audio: Triggering Audio Intelligence in Compact Language Models [28.38037427018435]
  We present Eureka-Audio, a compact yet high-performance audio language model that achieves competitive performance against larger models.<n>Despite containing only 1.7B parameters, Eureka-Audio demonstrates strong performance on automatic speech recognition (ASR), audio understanding, and dense audio captioning.<n>To further enhance paralinguistic reasoning, we introduce DataFlux, a closed loop audio instruction data synthesis and verification pipeline.
  arXiv  Detail & Related papers  (2026-02-15T02:01:08Z)
• DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment [94.0709779805955]
  We introduce DeSTA2.5-Audio, a general-purpose Large Audio Language Model (LALM)<n>It is designed for robust auditory perception and instruction-following, without requiring task-specific audio instruction-tuning.<n>DeSTA2.5-Audio achieves state-of-the-art or competitive performance across a wide range of audio-language benchmarks.
  arXiv  Detail & Related papers  (2025-07-03T16:28:25Z)
• From Alignment to Advancement: Bootstrapping Audio-Language Alignment with Synthetic Data [55.2480439325792]
  Audio-aware large language models (ALLMs) have recently made great strides in understanding and processing audio inputs.<n>These models are typically adapted from text-based large language models (LLMs) through additional training on audio-related tasks.<n>We propose a data generation framework that produces contrastive-like training data, designed to enhance ALLMs' ability to differentiate between present and absent sounds.
  arXiv  Detail & Related papers  (2025-05-26T16:08:41Z)
• VITA-Audio: Fast Interleaved Cross-Modal Token Generation for Efficient Large Speech-Language Model [70.25062476543091]
  VITA-Audio is an end-to-end large speech model with fast audio-text token generation.<n>MCTP module efficiently generates multiple audio tokens within a single model forward pass.<n>Four-stage progressive training strategy is explored to achieve model acceleration with minimal loss of speech quality.
  arXiv  Detail & Related papers  (2025-05-06T17:59:53Z)
• Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Models [95.45204813682885]
  We introduce Audio-Reasoner, a large-scale audio language model for deep reasoning in audio tasks.<n>We train Audio-Reasoner on CoTA, enabling it to achieve great logical capabilities in audio reasoning.<n>Our findings stress the core of structured CoT training in advancing audio reasoning.
  arXiv  Detail & Related papers  (2025-03-04T06:18:34Z)
• Multilingual Audio-Visual Speech Recognition with Hybrid CTC/RNN-T Fast Conformer [59.57249127943914]
  We present a multilingual Audio-Visual Speech Recognition model incorporating several enhancements to improve performance and audio noise robustness. We increase the amount of audio-visual training data for six distinct languages, generating automatic transcriptions of unlabelled multilingual datasets. Our proposed model achieves new state-of-the-art performance on the LRS3 dataset, reaching WER of 0.8%.
  arXiv  Detail & Related papers  (2024-03-14T01:16:32Z)
• Exploring the limits of decoder-only models trained on public speech recognition corpora [36.446905777292066]
  Decoder-Only Transformer for ASR (DOTA) model comprehensively outperforms the encoder-decoder open source replication of Whisper (OWSM) on nearly all English ASR benchmarks and outperforms Whisper large-v3 on 7 out of 15 test sets.
  arXiv  Detail & Related papers  (2024-01-31T23:29:42Z)
• tinyCLAP: Distilling Constrastive Language-Audio Pretrained Models [2.9619090219410515]
  This paper investigates how we can reduce the complexity of contrastive language-audio pre-trained models. We derive an unimodal distillation loss from first principles and explore how the dimensionality of the shared, multimodal latent space can be reduced. TinyCLAP uses only 6% of the original Microsoft CLAP parameters with a minimal reduction (less than 5%) in zero-shot classification performance.
  arXiv  Detail & Related papers  (2023-11-24T14:45:53Z)
• A Light Weight Model for Active Speaker Detection [7.253335671577093]
  We construct a lightweight active speaker detection architecture by reducing input candidates, splitting 2D and 3D convolutions for audio-visual feature extraction, and applying gated recurrent unit (GRU) with low computational complexity for cross-modal modeling. Experimental results on the AVA-ActiveSpeaker dataset show that our framework achieves competitive mAP performance (94.1% vs. 94.2%). Our framework also performs well on the Columbia dataset showing good robustness.
  arXiv  Detail & Related papers  (2023-03-08T08:40:56Z)
• Audio ALBERT: A Lite BERT for Self-supervised Learning of Audio Representation [51.37980448183019]
  We propose Audio ALBERT, a lite version of the self-supervised speech representation model. We show that Audio ALBERT is capable of achieving competitive performance with those huge models in the downstream tasks. In probing experiments, we find that the latent representations encode richer information of both phoneme and speaker than that of the last layer.
  arXiv  Detail & Related papers  (2020-05-18T10:42:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.