I am working on a project that requires the logit outputs (pre-softmax layer) on an accurate model for the CIFAR-10 dataset. I have been looking at Densenets since it has a softmax output layer. However, the current implementations that I've found either require training, or do not seem to be targeted towards CIFAR-10 (for instance the torchvision Densenets implmentation).
Are there any pretrained neural network models that are targeted for the CIFAR-10 dataset that would be easy to borrow and spit out the pre-softmax logit data? Preferably something like Densenets or as high accuracy as possible. Thanks!
