{"ID":2895525,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.08236","arxiv_id":"2507.08236","title":"Distilling Spectrograms into Tokens: Fast and Lightweight Bioacoustic Classification for BirdCLEF+ 2025","abstract":"The BirdCLEF+ 2025 challenge requires classifying 206 species, including birds, mammals, insects, and amphibians, from soundscape recordings under a strict 90-minute CPU-only inference deadline, making many state-of-the-art deep learning approaches impractical. To address this constraint, the DS@GT BirdCLEF team explored two strategies. First, we establish competitive baselines by optimizing pre-trained models from the Bioacoustics Model Zoo for CPU inference. Using TFLite, we achieved a nearly 10x inference speedup for the Perch model, enabling it to run in approximately 16 minutes and achieve a final ROC-AUC score of 0.729 on the public leaderboard post-competition and 0.711 on the private leaderboard. The best model from the zoo was BirdSetEfficientNetB1, with a public score of 0.810 and a private score of 0.778. Second, we introduce a novel, lightweight pipeline named Spectrogram Token Skip-Gram (STSG) that treats bioacoustics as a sequence modeling task. This method converts audio into discrete \"spectrogram tokens\" by clustering Mel-spectrograms using Faiss K-means and then learns high-quality contextual embeddings for these tokens in an unsupervised manner with a Word2Vec skip-gram model. For classification, embeddings within a 5-second window are averaged and passed to a linear model. With a projected inference time of 6 minutes for a 700-minute test set, the STSG approach achieved a final ROC-AUC public score of 0.559 and a private score of 0.520, demonstrating the viability of fast tokenization approaches with static embeddings for bioacoustic classification. Supporting code for this paper can be found at https://github.com/dsgt-arc/birdclef-2025.","short_abstract":"The BirdCLEF+ 2025 challenge requires classifying 206 species, including birds, mammals, insects, and amphibians, from soundscape recordings under a strict 90-minute CPU-only inference deadline, making many state-of-the-art deep learning approaches impractical. To address this constraint, the DS@GT BirdCLEF team explor...","url_abs":"https://arxiv.org/abs/2507.08236","url_pdf":"https://arxiv.org/pdf/2507.08236v1","authors":"[\"Anthony Miyaguchi\",\"Murilo Gustineli\",\"Adrian Cheung\"]","published":"2025-07-11T00:47:08Z","proceeding":"cs.SD","tasks":"[\"cs.SD\",\"eess.AS\"]","methods":"[]","has_code":false,"code_links":[{"ID":612202,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2895525,"paper_url":"https://arxiv.org/abs/2507.08236","paper_title":"Distilling Spectrograms into Tokens: Fast and Lightweight Bioacoustic Classification for BirdCLEF+ 2025","repo_url":"https://github.com/dsgt-arc/birdclef-2025","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}