Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Models
Paper
• 2402.19427 • Published
• 56
Curious
-
-
GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection
Paper • 2403.03507 • Published • 189 -
BiLLM: Pushing the Limit of Post-Training Quantization for LLMs
Paper • 2402.04291 • Published • 50 -
The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits
Paper • 2402.17764 • Published • 627