|
About the autograd category
|
|
0
|
3860
|
May 13, 2017
|
|
Missing argument create_graph in torch.func api
|
|
6
|
34
|
January 24, 2025
|
|
Backward pass error for loss computation in loop
|
|
1
|
7
|
January 23, 2025
|
|
Gradcheck fails for custom activation function
|
|
1
|
6
|
January 23, 2025
|
|
.grad should not equal None here
|
|
3
|
18
|
January 23, 2025
|
|
Unexpected behavior when using torch.autograd.functional.jacobian with multiple inputs/outputs neural network
|
|
3
|
30
|
January 21, 2025
|
|
Loss.backward() called after torch.nograd()
|
|
4
|
27
|
January 19, 2025
|
|
Where is the actual code for LayerNorm (torch.nn.functional.layer_norm)
|
|
6
|
3731
|
January 17, 2025
|
|
Free some saved tensors after partial backward
|
|
0
|
22
|
January 16, 2025
|
|
Why is requires_grad==False after multiplication?
|
|
9
|
388
|
January 16, 2025
|
|
Is it possible to have trainable module parameters in between static layer weights?
|
|
1
|
15
|
January 14, 2025
|
|
Torch.autograd.grad and masking issue
|
|
0
|
14
|
January 14, 2025
|
|
Unexpected behavior when using batch_jacobian with multiple inputs/outputs in quantum-classical neural network
|
|
2
|
31
|
January 12, 2025
|
|
Autograd FLOP Calculation with Higher Order Derivatives
|
|
0
|
22
|
January 11, 2025
|
|
Backward multiple forward passes
|
|
11
|
4257
|
January 11, 2025
|
|
Get a non-leaf tensor grad to its output
|
|
1
|
15
|
January 9, 2025
|
|
How to apply vmap on a heterogeneous tensor
|
|
5
|
58
|
January 6, 2025
|
|
How to figure out why my backwards pass is so slow?
|
|
2
|
28
|
January 6, 2025
|
|
How to remove backpropagation for specific tokens from the output of a transformer decoder only?
|
|
7
|
51
|
January 6, 2025
|
|
How does Pytorch Backward functions handle batched inputs?
|
|
4
|
72
|
January 4, 2025
|
|
Forward Mode AD with multiple tangents for each primal
|
|
1
|
21
|
January 3, 2025
|
|
Why the gradient values seems to be reversed in Tensor.backward()
|
|
9
|
19
|
January 2, 2025
|
|
Help with Projecting Gradients onto a Hypersphere's Tangent Plane
|
|
4
|
46
|
January 1, 2025
|
|
Autograd failing for sparse matmul at half precision
|
|
2
|
32
|
December 31, 2024
|
|
Custom layer-specific gradient calculation
|
|
1
|
38
|
December 31, 2024
|
|
How Does model.eval() Affect Gradient Descent in PyTorch and How to Handle Frequent Evaluations?
|
|
3
|
28
|
December 31, 2024
|
|
In-place Assignment of Leaf Tensors
|
|
1
|
49
|
December 29, 2024
|
|
Set parameters to model without breaking autograd
|
|
4
|
39
|
December 28, 2024
|
|
Composite RoPE backward gives a large ToCopyBackward0 in profiling trace
|
|
3
|
35
|
December 27, 2024
|
|
To detach or not to detach
|
|
1
|
74
|
December 26, 2024
|