DAQ: Decision-Aligned Post-Training Quantization for Spiking Transformers
arXiv:2611.22611v1 Announce Type: new Abstract: Spiking transformers combine attention-based representation learning with sparse spike communication, yet their weights and membrane variables still incur substantial storage and computation overhead. Post-training quantization (PTQ) offers a practical way to reduce these overheads, but preserving neuronal firing under weight and membrane quantization remains challenging. In this paper, we propose Decision-Aligned Post-Training Quantization (DAQ),
Read original sourceSource record
Project analysis
Project analysis is not yet available. See the source record above.