search

lxtGH / DecoupleSegNets

[ECCV-2020]: Improving Semantic Segmentation via Decoupled Body and Edge Supervision
370 stars 36 forks source link

How to train my own dataset? #3

Closed YBIO closed 4 years ago

YBIO commented 4 years ago

Hi, thank you for sharing your great work! I want to train the model on my own dataset (my dataset has 9 classes). Could you please give me some instructions? Thank u

lxtGH commented 4 years ago

@Shang-XH Thanks for your attention! 1, First you should write xxx.py (dataset) and write the get_item function according to your needs. 2, Then you should register your dataset in init.py by specifying your transform and loaders. Finaly add your data path in config.py.

I suggest you have a look at bdd.py to see how it works.

YBIO commented 4 years ago

@Shang-XH Thanks for your attention! 1, First you should write xxx.py (dataset) and write the get_item function according to your needs. 2, Then you should register your dataset in init.py by specifying your transform and loaders. Finaly add your data path in config.py.

I suggest you have a look at bdd.py to see how it works.

Thank u for the instruction~I will try it.