|
[SDPA] RTX5080 is different from CPU calculation result in backward with long seq
|
|
0
|
27
|
June 17, 2025
|
|
Function type in TorchScript
|
|
4
|
2516
|
June 17, 2025
|
|
TorchDispatchMode generates aten::detach for activations like ReLU
|
|
0
|
27
|
June 17, 2025
|
|
Building from GIT error: PyTorch needs CUDNN-8.5 or above but found 8.2.4
|
|
4
|
91
|
June 17, 2025
|
|
Using buffers in ParallelEnvs / MultiSyncCollectors
|
|
2
|
118
|
June 16, 2025
|
|
Wrongly transformed bounding box coordinates
|
|
4
|
68
|
June 16, 2025
|
|
Inconsistent Output for Identical Inputs When Using Linear Projection with Different squence length
|
|
4
|
75
|
June 16, 2025
|
|
[BUG] RTX5080： Function 'MmBackward0' returned nan values in its 0th output.
|
|
2
|
53
|
June 16, 2025
|
|
Ensuring same predict results cross Python and C++
|
|
2
|
86
|
June 16, 2025
|
|
Broken autograd momentum link
|
|
1
|
41
|
June 16, 2025
|
|
JVP and checkpointing
|
|
1
|
54
|
June 16, 2025
|
|
Scaled_dot_product_attention not useful for inference
|
|
0
|
40
|
June 15, 2025
|
|
Constant Predictions in Non-Linear Model Despite Training Progress
|
|
2
|
55
|
June 15, 2025
|
|
Loss.backward(): element 0 of tensors does not require grad and does not have a grad_fn
|
|
6
|
2451
|
June 15, 2025
|
|
Embed_dim must be divisible by num_heads
|
|
8
|
18692
|
June 15, 2025
|
|
Oom error during process in 3d side project
|
|
0
|
31
|
June 14, 2025
|
|
Multi-agent RL with different agent action spaces
|
|
0
|
44
|
June 12, 2025
|
|
Using an AWS S3 bucket dataset with Torchvision.Datasets.ImageFolder
|
|
0
|
39
|
June 13, 2025
|
|
How SGD works in pytorch
|
|
12
|
13723
|
March 17, 2023
|
|
RTX 5070Ti GPU and CUDA error
|
|
4
|
636
|
June 13, 2025
|
|
Custom autograd.Function for quantized C++ simulator
|
|
2
|
42
|
June 13, 2025
|
|
Windows & WSL2: zeroed CUDA tensors in spawned processes
|
|
0
|
43
|
June 13, 2025
|
|
What does cuda.is_initialized() actually check?
|
|
2
|
1232
|
June 13, 2025
|
|
Pytorch data loading best practices, any good resources to explore?
|
|
1
|
40
|
June 13, 2025
|
|
Evaluating gradients of output variables w.r.t parameters for pixelwise models
|
|
2
|
52
|
June 12, 2025
|
|
RuntimeError: expected scalar type Float but found BFloat16
|
|
1
|
109
|
June 12, 2025
|
|
Preserving bits like conj, neg during functionalization
|
|
0
|
17
|
June 12, 2025
|
|
Using a bidirectional nn.GRU Gated Recurrent Unit understand forwarding process
|
|
0
|
16
|
June 12, 2025
|
|
Cannot build PyTorch 2.3.1 with CUDA 12.9 on WSL2 (Ubuntu 22.04) — Missing nvToolsExt despite cuda-nvtx-12-9 installed
|
|
2
|
263
|
June 12, 2025
|
|
Simple "hello world" with torchtune on mac
|
|
1
|
49
|
June 12, 2025
|