Tips and Traps¶
- Adpative pooling an generate fixed-sized output, which enables a model to handle varying sized input. The models in torchvision are such examples.
References¶
Ways to Enable and Disable Gradient in PyTorch
set_grad_enable¶
Multiprocessing Issues in Docker
PyTorch Specific¶
Using num_workers > 0 for a DataLoader in Docker might causes issues ...
MULTIPROCESSING BEST PRACTICES
Runtime Error with DataLoader: exited unexpectedly #5301
https://forums.fast.ai/t/runtimeerror-dataloader-worker-pid-137-is-killed-by-signal-bus-error/27095
Get the Number of Cores in macOS
Using Shell Command¶
Delete a Layer in a Pretrained Model in PyTorch
It is common to customize a pretrained model by delete the output layer or replace it to the output layer that suits your use case. There are several ways to achieve this in PyTorch.
Replace the Fully Connected Layer with an Identity Layer¶
Define an identity layer.
Replace the fully connected layer with an Identity Layer (using ResNet18 as an example).
Save and Load PyTorch Models
Things on this page are fragmentary and immature notes/thoughts of the author. Please read with your own judgement!
-
PyTorch uses pickle to serialize and deserialize objects.
-
The PyTorch convention is to use the file extension
.ptor.pthfor saving model (or its parameters) and use the file extension …