LLM Pruning Comparison Hub
3 papers - avg viability 7.3
Top Papers
- High-Fidelity Pruning for Large Language Models(8.0)
Efficiently prune large language models using information entropy to reduce computational costs without sacrificing performance, offering a deployable solution for resource-constrained environments.
- Deterministic Differentiable Structured Pruning for Large Language Models(7.0)
DDP offers deterministic, differentiable pruning for LLMs, enabling faster inference with minimal performance loss, making it a valuable tool for optimizing LLM deployment.
- ROSE: Reordered SparseGPT for More Accurate One-Shot Large Language Models Pruning(7.0)
ROSE is a reordered SparseGPT method that prioritizes weights with larger potential pruning errors to be pruned earlier, leading to more accurate one-shot LLM pruning.