-
### 🐛 Describe the bug
Hello, we got `torch._dynamo.exc.Unsupported: Illegal getattr invocation requires_grad in strict mode` when trying to check the `requires_grad` attr of a tensor in backward. Ca…
-
### Describe the bug
Training a model with an LSTM module causes a segmentation fault after loss.backward() has been called a random number of times. The number of times the training loop can run …
-
### 🐛 Describe the bug
Hi!
I've trained EfficientNetB0 model from timm and converted it to .ptl with jit.trace. Here is my code:
```
dummy_input = torch.randn(1,3,256,256)
traced_script_module …
-
Status: Draft
Updated: 09/18/2024
# Objective
In this doc we’ll talk about how different optimization techniques are structured in torchao and how to contribute to torchao.
# torchao Stack Ove…
-
fft_r2c does not support dynamic shapes:
```python
import torch
class STFTModel(torch.nn.Module):
def __init__(self):
super().__init__()
self._window = torch.hann_windo…
-
## 🐛 Bug
Whenever I use `generate` function on a TPU (I use v5e litepod8), I have a crash with a C++ stack trace but no info on the python side and no way to catch/recover.
## To Reproduce
I…
-
# Requirement :
It's hard to debug rocMLIR issues right now because the trace variable is only effective during the final run, but not during the benchmarking/quick tuning stage, which is when we mi…
-
Hi @willow-ahrens @hameerabbasi,
This issue is meant to track progress of implementing Array API standard for `finch-tensor`.
I thought that we could try adding short notes to bullet-points, say…
-
问题描述:我参考infer/modules/onnx/export.py尝试将net_g的torch模型转onnx,遇到一些问题,望大佬帮忙看看:
1. onnx推理时不支持动态长度,export.py中默认hubert长度为200,转成onnx后也只能推理hubert长度为200的输入,其他长度的会报错;虽然问题通过简单的方式解决了,但是不知道会不会影响结果,我的解决方式是:根据转onnx的w…
-
When export dynamic-shape llama2, in the attached [piece of the partitioned model](https://github.com/user-attachments/files/16568296/partitioned-model_piece-2.txt), there are symbol manipulations suc…