Training function to generate NN weights


0

40

January 16, 2021

How could a BatchNorm layer generate nan in backward process?


1

57

January 16, 2021

Why is it so hard to enforce a weight matrix to be orthogonal?


4

68

January 16, 2021

Step Activation Function


3

738

March 3, 2020

RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` while running fine on the CPU


12

122

January 15, 2021

Equivalent of ignore_index in a continuous valued loss


0

28

January 15, 2021

Why it takes the same time to calculate the gradient of the last layer and all the gradients？


2

61

January 15, 2021

Using trained model for prediction with variable input batch size


1

35

January 15, 2021

Sharing a parameter between multiple loss functions/graphs


3

95

January 15, 2021

Gradients not computed after replacing BatchNorm with GroupNorm


0

23

January 14, 2021

CUDA error: an illegal memory access (when finetuning GroupNorm)


2

48

January 14, 2021

Curiously I don't get any gradients in one layer


2

33

January 14, 2021

Combine Losses and Weight those


8

64

January 14, 2021

Require gradient only for some tensor elements (others should be held constant)


9

568

January 14, 2021

How to implement torch.autograd.Function with arbitary length of input


1

31

January 13, 2021

Change model parameters before back propagation


3

75

January 13, 2021

How to save grad values after every batch?


2

38

January 12, 2021

Encountering error: Unrecognized tensor type ID: AutogradCUDA


6

285

January 12, 2021

Gradients exist but weights not updating


7

3427

January 12, 2021

Is it possible the edit the weight matrix after the forward pass but before the backward pass?


3

53

January 11, 2021

Functions with (almost) same output but (very) different gradients


0

28

January 11, 2021

Like TF2 tape style gradient


11

118

January 11, 2021

Use forward_pre_hook to modify nn.Module parameters


1

37

January 11, 2021

Adding new parameters for training


9

1127

January 11, 2021

Complex functions : "exp does not support automatic differentiation for outputs with complex dtype."


6

371

January 10, 2021

Optimizing regression weights for NN outputs


0

24

January 10, 2021

How to customize the double backward?


5

1145

January 9, 2021

How to compute Jacobian matrix in PyTorch?


15

13429

January 9, 2021

Adding scalar parameter


1

46

January 9, 2021

Lin's Concordance Correlation Coefficient as loss function


3

1278

January 9, 2021
