autograd


Topic Replies Activity
About the autograd category 1 May 13, 2017
Multiple submodule instances without merging gradients? 1 June 25, 2019
How can I wrap a parameter/tensor into a embedding matrix for look up 4 June 25, 2019
Runtime error, seems like graph is being destroyed 1 June 25, 2019
Quantization layer implementation 3 June 25, 2019
Submodel and requires_grad pb 1 June 24, 2019
What's the efficient way to lookup embedding from a given tensor 1 June 25, 2019
Difference between nn.MSELoss and torch.mean((op-target)**2) 2 June 24, 2019
Multi-task learning with adaptive weights for task losses 4 June 24, 2019
`torch.cuda.current_device()` always return 0 15 June 24, 2019
Explanation autograd, chain rule and Variables 2 June 24, 2019
Optimization of a logistic regression 1 June 24, 2019
How can test result change even when the learning rate is 0? 6 June 23, 2019
Warning: NaN or Inf found in input tensor 3 June 23, 2019
Question about how PyTorch handles Regularization 1 June 22, 2019
Batch Renormalization implementation in THCUNN 5 June 22, 2019
Parameter sharing along a dimension 1 June 22, 2019
RuntimeError: Function 'SqrtBackward' returned nan values in its 0th output 2 June 22, 2019
Backward for operations in torch.nn.functional 1 June 22, 2019
Changing the backward graph 1 June 22, 2019
Weird reults in training AlexNet 6 June 21, 2019
How to avoid overflow computing the graient of pow? 4 June 21, 2019
DataParallel Multi-GPU stuck on backward pass 1 June 21, 2019
How to Use a tensor as an Index to Parameter in Pytorch? 3 June 20, 2019
Stacking tensors 3 June 20, 2019
[nonissue] Autograd fails when using half-precision - overflow on matrix size 14 June 20, 2019
What is the problem with my code? 2 June 20, 2019
Training problem 2 June 20, 2019
Gradient descent back propagating for intermediate loss function 5 June 19, 2019
Can I manually edit/simplify the computational graph generated by autograd? 1 June 19, 2019