Unofficial PyTorch implementation of MobileViT v2. Most code is borrowed from official implement ml-cvnets
This project mean to make a friendly code for downstream task, like object detection, segmentation. etc. It's easy for integration. I had uploaded some converted model to BaiDuNetDisk for fine-tuning (convert from official pretrained model). I also compared inference result with original model and no suspicious result.
Model convert code in code tail of 'mobilevit_v2.py', also can directly download those model from BaiDuNetDisk
Pretrain Model Link extract code [6666]