相关数据集
MLDS
MLDS是一个由马里兰大学巴尔的摩分校的John Clemens创建的大型数据集,包含超过22万个经过训练的神经网络模型。这些模型通过MLC@Home项目,一个基于全球志愿者分布式计算的平台生成。数据集旨在通过直接分析神经网络的结构和权重,解决传统评估方法如损失函数难以检测的问题。MLDS数据集不仅用于模型间的比较,还用于研究模型与训练数据之间的关系,特别是在模型安全性和可追溯性方面的应用。
arXiv2021-04-21 更新470
NeuralNetwork
Multi-fidelity tabular data for Neural Network hyperparameter space for HPOBench
DataCite Commons2022-05-30 更新80
adc2025-training-nn-models-add-awp-ver7-models
https://www.kaggle.com/code/horikitasaku/adc2025-test-submission-change-ver7
kaggle2025-09-19 更新60
Additional file 4: Table S3. of Medical implications of technical accuracy in genome sequencing
Sites with systematic errors prior to filtering that are present in ClinVar. (XLSX 40 kb)
DataCite Commons2024-12-18 更新100
Benchmark synthesis
A dataset of literature-based and custom neural networks, for the purpose of testing the performance of trained models at predicting the required resources and inference latency on FPGA.
DataCite Commons2025-11-20 更新70



