ComplexFormer: Disruptively Advancing Transformer Inference Ability via Head-Specific Complex Vector Attention

Documenti analoghi

• VQ-Logits: Compressing the Output Bottleneck of Large Language Models via Vector Quantized Logits
  di: Shao, Jintian, et al.
  Pubblicazione: (2025)
• Towards Analyzing and Understanding the Limitations of VAPO: A Theoretical Perspective
  di: Shao, Jintian, et al.
  Pubblicazione: (2025)
• EulerFormer: Sequential User Behavior Modeling with Complex Vector Attention
  di: Tian, Zhen, et al.
  Pubblicazione: (2024)
• CipherFormer: Efficient Transformer Private Inference with Low Round Complexity
  di: Wang, Weize, et al.
  Pubblicazione: (2024)
• RecurFormer: Not All Transformer Heads Need Self-Attention
  di: Yan, Ruiqing, et al.
  Pubblicazione: (2024)