AI Metadata Enrichment Artifacts of GPU-Server
收藏Zenodo2026-06-20 更新2026-06-21 收录
下载链接：
https://zenodo.org/doi/10.5281/zenodo.20765512
下载链接
链接失效反馈官方服务：
资源简介：
Contains the AI metadata enrichment pipeline artifacts produced on the GPU-Server for three RO-Crate artifacts:
MoonGen Load Latency Test -> 2026-06-18_18-40-22_785154-moongen
QUIC Implementation Benchmark -> 2026-06-18_20-34-16_305947-quic
Multipath Topology Experiment -> 2026-06-18_22-42-17_403726-multipath
Each experiment run enriches the specific RO-Crate artifact with generated titles, descriptions, and keywords using locally hosted LLMs through Ollama. Different models and prompt context budgets are combined to evaluate how model choice and available artifact context influence metadata quality. Each run produces an enriched RO-Crate artifact together with run manifests, including token and runtime statistics, GPU and energy measurements, and automated comparisons against a human-written reference using BERTScore metric.
Browse the artifacts here:
https://tumi8.github.io/paper-llm-artifacts-ai-enrichments-gpu-server/
提供机构：
Zenodo创建时间：
2026-06-19
