MaybeShewill-CV / bisenetv2-tensorflow

Unofficial tensorflow implementation of real-time scene image segmentation model "BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation"
https://maybeshewill-cv.github.io/bisenetv2-tensorflow/
MIT License
224 stars 59 forks source link

请问怎么使用使用 v2 large方式训练 #42

Closed dgliu123 closed 3 years ago

dgliu123 commented 3 years ago

请问怎么使用使用 v2 large方式训练

MaybeShewill-CV commented 3 years ago

@dgliu123 这个目前还是得通过改代码的方式==!

dgliu123 commented 3 years ago

我不理解 v2 large方式是个什么方式

MaybeShewill-CV commented 3 years ago

@dgliu123 网络结构和现在略有不同 增加了广度和深度 具体你可以看下论文里的描述:)