Search
❯
Jun 17, 20261 min read
A way to initialize model's lazy layers.
model(torch.zeros([*input_shape]))