MUXQ: Mixed-to-Uniform Precision MatriX Quantization via Low-Rank Outlier Decomposition

Documenti analoghi

• Assigning Distinct Roles to Quantized and Low-Rank Matrices Toward Optimal Weight Decomposition
  di: Cho, Yoonjun, et al.
  Pubblicazione: (2025)
• OATS: Outlier-Aware Pruning Through Sparse and Low Rank Decomposition
  di: Zhang, Stephen, et al.
  Pubblicazione: (2024)
• Towards Scalable Handwriting Communication via EEG Decoding and Latent Embedding Integration
  di: Kim, Jun-Young, et al.
  Pubblicazione: (2024)
• FedWSQ: Efficient Federated Learning with Weight Standardization and Distribution-Aware Non-Uniform Quantization
  di: Kim, Seung-Wook, et al.
  Pubblicazione: (2025)
• STaMP: Sequence Transformation and Mixed Precision for Low-Precision Activation Quantization
  di: Federici, Marco, et al.
  Pubblicazione: (2025)