I-Segmenter: Integer-Only Vision Transformer for Efficient Semantic Segmentation
Vision Transformers (ViTs) have recently achieved strong results in semantic segmentation, yet their deployment on resource-constrained devices remains limited due to their high memory footprint and computational cost. Quantization offers an effective strategy to improve efficiency, but ViT-based segmentation models are notoriously fragile under low precision, as quantization errors accumulate across deep encoder-decoder pipelines. We introduce I-Segmenter, the first fully integer-only ViT segmentation framework. Building on the Segmenter architecture, I-Segmenter systematically replaces floating-point operations with integer-only counterparts. To further stabilize both training and inference, we propose $λ$-ShiftGELU, a novel activation function that mitigates the limitations of uniform quantization in handling long-tailed activation distributions. In addition, we remove the L2 normalization layer and replace bilinear interpolation in the decoder with nearest neighbor upsampling, ensuring integer-only execution throughout the computational graph. Extensive experiments show that I-Segmenter achieves accuracy within a reasonable margin of its FP32 baseline (5.1 % on average), while reducing model size by up to 3.8x and enabling up to 1.2x faster inference with optimized runtimes. Notably, even in one-shot PTQ with a single calibration image, I-Segmenter delivers competitive accuracy, underscoring its practicality for real-world deployment.
Code (0)
등록된 구현이 없습니다.
Tasks
Semantic SegmentationSimilar Papers 제목 키워드 기반
Segmenter: Transformer for Semantic Segmentation
Image segmentation is often ambiguous at the level of individual image patches and requires contextual information to reach label consensus. In this paper we introduce Segmenter, a transformer model for semantic segmenta…
Decoderimage-classificationImage ClassificationImage Segmentation+3From Semantic Segmentation of Natural Images to Medical Image Segmentation Using ViT-Based Architectures
We address the problem of medical image segmentation in the context of limited training data. Our approach builds on the capabilities of the Vision Transformer (ViT) and the recent Segmenter model, adapting them for the …
Image SegmentationLesion SegmentationMedical Image SegmentationSemantic SegmentationGraph-Segmenter: Graph Transformer with Boundary-aware Attention for Semantic Segmentation
The transformer-based semantic segmentation approaches, which divide the image into different regions by sliding windows and model the relation inside each window, have achieved outstanding success. However, since the re…
RelationSegmentationSemantic SegmentationIPTQ-ViT: Post-Training Quantization of Non-linear Functions for Integer-only Vision Transformers
Previous Quantization-Aware Training (QAT) methods for vision transformers rely on expensive retraining to recover accuracy loss in non-linear layer quantization, limiting their use in resource-constrained environments. …
Image ClassificationObject DetectionI-ViT: Integer-only Quantization for Efficient Vision Transformer Inference
Vision Transformers (ViTs) have achieved state-of-the-art performance on various computer vision applications. However, these models have considerable storage and computational overheads, making their deployment and effi…
Quantization