|
Runtime error raised in DDP when using .detach() to skip gradient computation in some DP ranks
|
|
2
|
45
|
January 28, 2026
|
|
FSDP2 vs DDP gradient mismatch on Embeddings (Flex Attention + Compile)
|
|
0
|
42
|
January 27, 2026
|
|
Huge accuracy drop from QAT model after convert_pt2e
|
|
1
|
45
|
January 27, 2026
|
|
[Distributed w/ TorchTitan] Introducing Async Tensor Parallelism in PyTorch
|
|
12
|
17708
|
January 27, 2026
|
|
Multi GPU training on single node with DistributedDataParallel
|
|
3
|
5441
|
January 27, 2026
|
|
Pytorch DataLoader vs Tensorflow TFRecord
|
|
5
|
16591
|
January 25, 2026
|
|
Best practices for network bottlenecked image data loading + transforms
|
|
2
|
64
|
January 23, 2026
|
|
Model output explode and causing NaN loss after training for a few steps
|
|
1
|
41
|
January 23, 2026
|
|
Set seed doesn't work
|
|
3
|
53
|
January 23, 2026
|
|
Torch.prod produces RuntimeError: CUDA driver error: invalid argument
|
|
30
|
9370
|
January 22, 2026
|
|
Custom ops library with new type of neuron for PyTorch
|
|
0
|
27
|
January 22, 2026
|
|
PyTorch Day India 2026, Bengaluru – Anyone attending?
|
|
1
|
61
|
January 22, 2026
|
|
Torch.compile on train_step incl both fwd and bwd
|
|
1
|
39
|
January 21, 2026
|
|
MacBook Air M4 Chip PyTorch Compatability
|
|
1
|
199
|
January 21, 2026
|
|
Uninstall PyTorch completely to install older version
|
|
11
|
2948
|
January 21, 2026
|
|
Parallelizing matrix-matrix-matrix products
|
|
2
|
30
|
January 21, 2026
|
|
Fast symmetric matrix per vector multiplication
|
|
2
|
34
|
January 21, 2026
|
|
Out Of Memory Error CUDA
|
|
5
|
12571
|
January 20, 2026
|
|
8xH100 training issue
|
|
4
|
124
|
January 20, 2026
|
|
DGX Spark GB10 Cuda 13.0 Python 3.12 SM_121
|
|
16
|
2003
|
January 20, 2026
|
|
AOTInductor on Windows
|
|
0
|
51
|
January 19, 2026
|
|
Cuda and Torch install via pip vs. conda
|
|
5
|
178
|
January 19, 2026
|
|
It seems Pytorch doesn't use GPU
|
|
12
|
26125
|
January 18, 2026
|
|
Torch.compile Diagnostic Dashboard
|
|
1
|
49
|
January 18, 2026
|
|
Varying batch size for pre-trained model changes inference result, even in evaluation mode
|
|
2
|
74
|
January 18, 2026
|
|
ResDAG: A modern, GPU-accelerated reservoir computing library for PyTorch
|
|
0
|
76
|
January 18, 2026
|
|
Pytorch Symbolic: an equivalent of Keras Functional API
|
|
3
|
1802
|
January 18, 2026
|
|
User warning about `enable_nested_tensor`
|
|
6
|
3382
|
January 18, 2026
|
|
Difference between two tensors using the same device
|
|
4
|
36
|
January 18, 2026
|
|
[WinError 1114] loading c10.dll issue occurred while running import torch
|
|
1
|
901
|
January 17, 2026
|