samsungsailmontreal / ghn3 Goto Github PK

View Code? Open in Web Editor NEW

28.0 4.0 3.0 1.2 MB

Code for "Can We Scale Transformers to Predict Parameters of Diverse ImageNet Models?" [ICML 2023]

Home Page: https://arxiv.org/abs/2303.04143

License: MIT License

Python 8.57% Shell 91.43%

hypernetworks imagenet large-scale transformers computational-graphs graphs deep-learning pytorch

ghn3's Issues

Need additional detail regarding few-shot learning experiment

Can you explain the few-shot setting whose result are reported in
Table 7. Transfer learning from ImageNet to few-shot CIFAR-10
and CIFAR-100 with 1000 training labels with 3 networks: ResNet50 (R-50), ConvNext-B (C-B) and Swin-T (S-T)

How many shots in the training set as well as in the test set or query set.

Questions about the paper

Thank you for your interesting research. I have some questions regarding the paper:

I'm curious about the adaptability of GHNs to other standard-sized datasets, particularly in different tasks such as image segmentation. The Penn-Fudan dataset discussed in your paper seems relatively small. Could you share your thoughts on this?
Do you remember the accuracy of the models in DeepNets-1M achieved during the training of GHNs model? How are they compared to the predicted parameter model accuracy with/without fine-tuning?

Thank you.

Recommend Projects