How to make a Deep Learning Project Structure?
- model configs
- Choose the format
.json
.ini
.yaml
.py
- your datasets
- train
- class 1
- class 2
- etc.
- valid
- class 1
- class 2
- etc.
- test
- class 1
- class 2
- etc.
- train
- project documents
- install
- how to run?
- api structure
- save sample image
- dataloader output image (+ transforms)
- sample image for test
- result
- your trained model
- data processing notebook
- tutorial notebook
- train scripts
- test scripts
- run scripts
-
/data
- dataloader
- data utils
-
/model
- model architecture
- model utils
-
/visualization
- visualization tools
-
loss.py
and/loss
- define custom loss
-
optim.py
and/optim
- define custom optimizer
-
scheduler.py
and/scheduler
- define custom scheduler
-
transforms.py
and/transforms
- define custom data augmentation