PRATIK MAZUMDER

FPAIT - How are fully connected layers generating 3D weights for Convolutional layer outputs?

7

Hi, could you please explain how the fully connected layers are being used to convert the 1 dimensional embedding of the text into a mutli-dimensional weight matrix for the convolutional...

Hi, how many epochs did you train on the CIFAR for the results reported for ALI and ALICE. Was it 100 as given in the code or more?

test_semisup gives Validation error rate around 0.12 for svhn after training for 100 epochs

8

Which configuration is this referring to in the original paper for SVHN test set misclassification rate. Is it: 1) ALI (ours, L2-SVM) ----->19.14 ± 0.50 (in paper) or 2) ALI...

Confusion Regarding belga2toplogo result

8

I got the following result in the file **best_iou.txt** for **belga2toplogo** (main_train_val_test.py) which doesn't seem to match the reported accuracy in the paper: Best score epoch: 100 acc cls: 0.6151...

How many epochs does it take roughly to converge for gtsrb2TT100K?

1

Hi, How many epochs does it take roughly before the results start matching the reported results for gtsrb2TT100K

Hi @gidariss, Can you please provide the config files for the other Networks. For example, for mini-imagenet, only the wideresnet config files have been provided. Thanks in advance

Cannot reproduce results

1

Hi, I ran the experiments for colored MNIST three times and got the following results: python train.py with server_user colored_mnist skewed3 severity4 valid_accs_d aligned 0.5823 valid_accs_d skewed 0.7024 valid_accs_d aligned...