⚡️Inspiration: Advice to swanlab.init

[Zeyi-Lin]

For example, the following code:

# Initialize swanlab
run = swanlab.init(
    config={
        "lr": 1e-4,
        "num_epochs": 40,
    }
)

optimizer = torch.optim.Adam(model.parameters(), lr=run.lr)

for epoch in range(run.num_epochs):
    ···

run is an instance after the initialization of swanlab, so the content in config can be called in the form of run.lr or run.num_epochs.

I think this can reduce some repetitive code when writing machine learning code to improve efficiency.

[Zeyi-Lin]

v0.1.4