Please enable JavaScript.
Coggle requires JavaScript to display documents.
Pytorch Going Modular - Coggle Diagram
Pytorch Going Modular
- Train, evaluate and save the model (train.py)
-
-
-
- Setup various hyperparameters
-
-
-
-
-
-
-
the scripts: going_modular directory, data_setup, engine, model_builder, utils.
dependencies, namely: torch, os, torchvision.transforms
-
-
-
-
- Creating train_step() and test_step() functions and train() to combine them --- (engine.py())
def test_step(model,dataLoader,loss_fn,device)
-
def train(model,train_dataloader,test_dataloader,optimizer,loss_fn,epoch,device)
-
def train_step(model,dataLoader,loss_fn,optimizer,device)
-
- Cell mode vs. script mode
- Going Modular: Part 1 (cell mode)
-
- Going Modular: Part 2 (script mode)
-
- Create Datasets and DataLoaders (data_setup.py)
-
- Making a model (model_builder.py)
-
- Creating a function to save the model (utils.py)
def save_model(model,target_dir,model_name)