Project 2 2026 Trained Model Weights
This repository contains the PyTorch checkpoints required by the final report
reports/report_submit.md for Project 2 of "Neural Network and Deep Learning".
Dataset: CIFAR-10 official dataset, https://www.cs.toronto.edu/~kriz/cifar.html
Code: https://github.com/TimoZhou1024/pj2_nndl
Included Checkpoints
The uploaded files preserve the report-relative paths under
reports/experiments/.../weights/....
Task 1 CIFAR-10 model-search checkpoints:
reports/experiments/task1/basic_cnn_baseline/weights/best_basic_cnn.ptreports/experiments/task1/vgg_a_light_baseline/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_width_0_5/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_width_0_75/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_width_1_5/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_width_2_0/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_width_2_5/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_label_smoothing/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_cross_entropy_wd_1e_4/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_multi_margin_wd_1e_4/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_label_smoothing_leaky_relu/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_label_smoothing_elu/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_label_smoothing_leaky_relu_adamw_wd_1e_4/weights/best_vgg_a_light.ptreports/experiments/task1/vgg_a_light_w2_label_smoothing_leaky_relu_sgd_momentum_lr_1e_2/weights/best_vgg_a_light.pt
Task 2 VGG-A and VGG-A+BatchNorm learning-rate sweep checkpoints:
reports/experiments/task2/task2_vgg_a_lr_1e_3/weights/best_vgg_a.ptreports/experiments/task2/task2_vgg_a_lr_2e_3/weights/best_vgg_a.ptreports/experiments/task2/task2_vgg_a_lr_5e_4/weights/best_vgg_a.ptreports/experiments/task2/task2_vgg_a_lr_1e_4/weights/best_vgg_a.ptreports/experiments/task2/task2_vgg_batchnorm_lr_1e_3/weights/best_vgg_batchnorm.ptreports/experiments/task2/task2_vgg_batchnorm_lr_2e_3/weights/best_vgg_batchnorm.ptreports/experiments/task2/task2_vgg_batchnorm_lr_5e_4/weights/best_vgg_batchnorm.ptreports/experiments/task2/task2_vgg_batchnorm_lr_1e_4/weights/best_vgg_batchnorm.pt
Supplementary Task 2 full-gradient rerun checkpoints:
reports/experiments/task2_fullgrad/task2_fullgrad_vgg_a_lr_1e_3/weights/best_vgg_a.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_a_lr_2e_3/weights/best_vgg_a.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_a_lr_5e_4/weights/best_vgg_a.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_a_lr_1e_4/weights/best_vgg_a.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_batchnorm_lr_1e_3/weights/best_vgg_batchnorm.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_batchnorm_lr_2e_3/weights/best_vgg_batchnorm.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_batchnorm_lr_5e_4/weights/best_vgg_batchnorm.ptreports/experiments/task2_fullgrad/task2_fullgrad_vgg_batchnorm_lr_1e_4/weights/best_vgg_batchnorm.pt