Problemsby IbisTrouble with PyTorch eager
MLOps2 posts from 2 people2d active+00 posts in the last 7 days, 1 the 7 days before (fading)Unverified
Posts per day
The posts behind it
2, newest first`torch.compile channel_shuffle identity shortcut preserves non-contiguous strides unlike eager CUDA, causing downstream view to fail`### 🐛 Describe the bug A non-contiguous CUDA tensor passed through `torch.channel_shuffle(x, 1)` and then a legal `.view(...)` works in eager but crashes under `torch.compile` on **all** backends (`eager`, `aot_eager`, `inductor`) during fake-tensor propagation. Since...pytorch/pytorch · tinywisdomOpen ↗F.threshold on integer tensors truncates a fractional threshold to the tensor dtype; eager and torch.compile disagree### 🐛 Describe the bug `torch.nn.functional.threshold(input, threshold, value)` is documented as `y = x if x > threshold else value`. For an integer `input` and a fractional `threshold`, the kernels cast `threshold` to the input dtype (`threshold_scalar.to ()` in...pytorch/pytorch · the-Shallow1 repliesOpen ↗
Companies and products named
About this problem
Evidence
2 posts from 2 people in 1 place, about 2 a week over 2 days. Mostly on pytorch/pytorch. Tools named alongside: NVIDIA.
Frustration Frustration 0 of 3· Seen on GitHub issues
How it was grouped
Posts about PyTorch with no known issue, sharing the word "eager". First post Oct 7, 2026, latest Oct 8, 2026. Unverified: fewer than 3 posts, or one person or place only. Method
Other PyTorch problems
PyTorch users keep filing bugsTrouble with PyTorch inputTrouble with PyTorch truePyTorch deprecations force unplanned migrationsPyTorch upgrades break working setupsTrouble with PyTorch workerPyTorch builds are slow or flakyTrouble with PyTorch inductorPyTorch GPU access and drivers are painfulTrouble with PyTorch testPyTorch vulnerabilities are hard to triage and patchRunning PyTorch locally hits memory and speed limits
History
steady to fading
new to steady
Trouble with PyTorch eager (2 posts)