search

keyu-tian / SparK

[ICLR'23 Spotlight🔥] The first successful BERT/MAE-style pretraining on any convolutional network; Pytorch impl. of "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"
https://arxiv.org/abs/2301.03580
MIT License
1.42k stars 82 forks source link

pretrained model on imagenet #5

Closed AMNAALMGLY closed 1 year ago

AMNAALMGLY commented 1 year ago

the link to pretrained imagenet model is broken?

keyu-tian commented 1 year ago

We now finish them 🤗, please check: https://github.com/keyu-tian/SparK#imagenet-1k-results-and-pre-trained-checkpoint-files.

Also we're working on uploading these weights to timm and huggingface, once it finish we would come back to you.