Architecture Descriptions and High Frequency Accuracy and Loss Data of Random Neural Networks Trained on Image Datasets

Contents & fields

This dataset contains architecture descriptions and high-frequency accuracy and loss data of random neural networks. Each network is trained for 40 epochs, with training loss, validation loss, and validation accuracy recorded every half epoch. All generated parameter values are recorded to allow network reconstruction. Data is stored in tables, one row per network. No field-level description is publicly available on the source page.

Research uses

Can be used for early prediction methods in neural architecture search (NAS), analyzing loss and accuracy curves, and comparing network performance with different parameters.

Information comes from the source page. Please check that page for current details and terms.