-
Hello,
When I execute ```torch._export.capture_pre_autograd_graph(model, inputs)``` , The the following code throws an error
```
torch.nn.utils.weight_norm(
ConvTranspose1d(
…
-
I'm looking at some models in MONAI which involves `torch.nn.Upsample`. I notice that torchscript exports the Upsample module to a `Resize` node but dynamo exports it to a very big graph and has a per…
-
Would be nice if we could configure the output rate of the Frost Dynamo and stat boosts each augment gives.
KubeJS support for the Frost Dynamo would be great as well.
I also found that the Fros…
-
### 🐛 Describe the bug
Building on Debian 10 using the default compiler GCC 8.3 fails. It looks that the [[likely]] and [[unlikely]] attributes are not well supported by this GCC version.
### Error …
-
### 🐛 Describe the bug
### Reproducer
```python
import torch
@torch.compile(backend="aot_eager", fullgraph=True)
def fn(x):
return x.sin()
x = torch.randn(3, 4)
y = torch.func.vm…
-
### Dynamo Version
3.3.0.5365
### Host
None
### Operating System
Microsoft Windows 11
### What did you do?
Continuous Undo
### What did you expect to see?
Previously deleted codeblock
### Wh…
-
### 🐛 Describe the bug
I use different Systems: `Python 3.11, torch 2.2.2, Ubuntu 22.04 or 23.10 (x64), all Cuda 12.x, all Intel CPUs (11..13 gen), all using pip not conda` system language setting …
-
### 🐛 Describe the bug
torchbench_amp_fp16_training
xpu train timm_regnet
E0626 18:18:36.100000 139652021139264 torch/_dynamo/utils.py:1478] RMSE (res-fp64): 0.00227, (ref-fp64): 0.00064 and sha…
-
### Dynamo Version
Dynamo Core: 2.16.2.5624
Dynamo Revit: 2.16.3.11903
### Revit Version
Revit 2023
### Operating System
Windows 10
### What did you do?
I use a few python scripts, which also u…
-
# Issue Description
Please fill in the following information to help us reproduce the issue:
## Dynamo version
Dynamo: `2.13.1.3887`
## Operating system
OS: `Microsoft Windows NT 10.0.19045.0…