|
Post training quantized model gets the error "Copying from quantized Tensor to non-quantized Tensor is not allowed" even though I'm not copying tensor
|
|
5
|
44
|
February 12, 2026
|
|
Mark_dynamic does not work when the dimension is 1
|
|
0
|
5
|
February 12, 2026
|
|
Pytorch dev 130
|
|
1
|
11
|
February 12, 2026
|
|
Is torch Muon optimizer compatible with FSDP/HSDP?
|
|
1
|
34
|
February 12, 2026
|
|
Is there any way to compute multiple jvps without repeatly computing function value?
|
|
4
|
28
|
February 12, 2026
|
|
Flex_attention and torch.compile
|
|
1
|
11
|
February 12, 2026
|
|
Question About Backward–ReduceScatter Overlap in FSDP Figure 5
|
|
0
|
7
|
February 12, 2026
|
|
Unable to allocate shared memory(shm) for file
|
|
0
|
7
|
February 12, 2026
|
|
Cuda Memory Use Batched Matrix Multiplication
|
|
7
|
2044
|
February 11, 2026
|
|
ONNX Community Survey 2026
|
|
0
|
10
|
February 11, 2026
|
|
Help training Titan+MIRAS, model learns to cheat loss
|
|
0
|
8
|
February 11, 2026
|
|
Is it safe to create new gpu tensor during cuda graph capturing?
|
|
3
|
36
|
February 10, 2026
|
|
Looking for matmul.h
|
|
1
|
22
|
February 10, 2026
|
|
"Offset increment outside graph capture encountered unexpectedly" is a specific PyTorch Windows issue that's known to occur with CUDA Graphs. This is a bug in PyTorch Windows builds
|
|
0
|
11
|
February 10, 2026
|
|
DEBUG=1 build with NDEBUG flag possible? (CFLAGS override related)
|
|
0
|
17
|
February 10, 2026
|
|
Reduce_sum() or sum operation (reduction operation) for floating point ops isnt accurate or not implementing kahan summation algorithm ,instead just returning a+b --->leading to vanishing gradient problem
|
|
1
|
25
|
February 10, 2026
|
|
Severe training slowdown after unfold-based patchification
|
|
4
|
38
|
February 10, 2026
|
|
Request: KV Cache Steering for VLM Hallucination Mitigation
|
|
0
|
11
|
February 9, 2026
|
|
DP with opacus: Want to understanding the functions
|
|
2
|
135
|
February 9, 2026
|
|
Function 'Scaled Dot Product Efficient Attention Backward0' returned nan values in its 0th output
|
|
13
|
2099
|
February 9, 2026
|
|
Can 'torch.profiler' support GB20X NVIDIA GPU architectures?
|
|
2
|
32
|
February 8, 2026
|
|
Torch.export does not support specify partial output
|
|
2
|
24
|
February 8, 2026
|
|
Diskoffloading during reverse mode
|
|
0
|
15
|
February 8, 2026
|
|
NaN Loss Issues with Precision 16 in PyTorch Lightning GAN Training
|
|
9
|
3958
|
February 6, 2026
|
|
[ROCm][CI] fp8 acceptable accuracy threshold
|
|
0
|
35
|
February 6, 2026
|
|
RNN memory management: RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation
|
|
2
|
26
|
February 6, 2026
|
|
Pretrained dendritic ResNet-18: 4x more parameter-efficient than ResNet-34
|
|
0
|
19
|
February 6, 2026
|
|
Pytorch not compatible with rtx 5050
|
|
5
|
85
|
February 6, 2026
|
|
Multi-GPU training with training loop that can skip backpropagation
|
|
1
|
130
|
February 6, 2026
|
|
At / tensor indexing helper for stable API
|
|
0
|
23
|
February 5, 2026
|