Submitted by akhaliq 44 QA-LoRA: Quantization-Aware Low-Rank Adaptation of Large Language Models · 9 authors 8
Submitted by akhaliq 42 LAVIE: High-Quality Video Generation with Cascaded Latent Diffusion Models · 20 authors 3
Submitted by akhaliq 18 DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models · 7 authors 1
Submitted by akhaliq 7 Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Models · 8 authors 1