Softmax Reparameterization for Output-Head Quantization
Paper • 2609.31291 • Published • 6
Shangzhen Zhu
ArgonnSZ
·
AI & ML interests
None yet
Recent Activity
submitted a paper 15 days ago
Pretraining Transformers with Quantized Softmax in Attention upvoted a paper 15 days ago
Softmax Reparameterization for Output-Head Quantization upvoted a paper 15 days ago
Approximating Softmax in Pretrained LLMs: Model Sensitivity and Kernel AccelerationOrganizations
