| Instantiating a big model | |
| When you want to use a very big pretrained model, one challenge is to minimize the use of the RAM. The usual workflow | |
| from PyTorch is: | |
| Create your model with random weights. | |
| Load your pretrained weights. | |
| Put those pretrained weights in your random model. | |
| Step 1 and 2 both require a full version of the model in memory, which is not a problem in most cases, but if your model starts weighing several GigaBytes, those two copies can make you get out of RAM. |