FastFusionNet: New State-of-the-Art for DAWNBench SQuAD
Felix Wu, Boyi Li, Lequn Wang, Ni Lao, John Blitzer, Kilian Q. Weinberger
Abstract
In this technical report, we introduce FastFusionNet, an efficient variant of FusionNet [12]. FusionNet is a high performing reading comprehension architecture, which was designed primarily for maximum retrieval accuracy with less regard towards computational requirements. For FastFusionNets we remove the expensive CoVe layers [21] and substitute the BiLSTMs with far more efficient SRU layers [19]. The resulting architecture obtains state-of-the-art results on DAWNBench [5] while achieving the lowest training and inference time on SQuAD [25] to-date. The code is available at https://github.com/felixgwu/FastFusionNet.
Related Papers
From Roots to Rewards: Dynamic Tree Reasoning with RL2025-07-17HapticCap: A Multimodal Dataset and Task for Understanding User Experience of Vibration Haptic Signals2025-07-17A Survey of Context Engineering for Large Language Models2025-07-17MCoT-RE: Multi-Faceted Chain-of-Thought and Re-Ranking for Training-Free Zero-Shot Composed Image Retrieval2025-07-17Developing Visual Augmented Q&A System using Scalable Vision Embedding Retrieval & Late Interaction Re-ranker2025-07-16Language-Guided Contrastive Audio-Visual Masked Autoencoder with Automatically Generated Audio-Visual-Text Triplets from Videos2025-07-16Context-Aware Search and Retrieval Over Erasure Channels2025-07-16Seq vs Seq: An Open Suite of Paired Encoders and Decoders2025-07-15