Tag
The paper proposes REP-LIE, a resource-efficient pruning method for Transformer models that uses low-rank importance estimation from LoRA gradients to enable pruning during finetuning, with competitive performance on LLaMA-7B and Mistral-7B.
The author introduces Capstan, a lightweight coding agent with a C core and hot-reloadable Lua plugins, designed for efficiency and extensibility in development workflows.