search

HUANGLIZI / LViT

[IEEE Transactions on Medical Imaging/TMI] This repo is the official implementation of "LViT: Language meets Vision Transformer in Medical Image Segmentation"
MIT License
283 stars 26 forks source link

Pre-training #42

Closed acmername closed 5 months ago

acmername commented 5 months ago

Do the experimental results obtained in Table 2 of the paper use a pre-trained model? Using pre-training works well on the COVID-19 dataset, but the other one is not good.If not used, the opposite experiment is obtained.

HUANGLIZI commented 5 months ago

The performance may have something with the GPU used and the hyperparameters may be different. So you can choose what is best for your setup.