yaoshiang
yaoshiang
### 🚀 The feature, motivation and pitch This feature would allow developers to fine tune on smaller GPUs and / or larger batch sizes, likely leading to higher MFU. Currently,...
### System Info (llama-dev) azureuser@yh-a100:~/cloudfiles/code/test_llama$ python -m "torch.utils.collect_env" /anaconda/envs/llama-dev/lib/python3.10/runpy.py:126: RuntimeWarning: 'torch.utils.collect_env' found in sys.modules after import of package 'torch.utils', but prior to execution of 'torch.utils.collect_env'; this may result in unpredictable...
# What does this PR do? Makes a minor fix to the parsing of the `--custom_dataset.file` flag. The documentation says you can add a colon in this value to specify...
### 🐛 Describe the bug CelebA bounding boxes were calculated on the so called "in-the-wild" images, prior to cropping and resizing. But torchvision.datasets returns the version that is cropped to...
there are five instances of code like this in the repo: ```python if type(v) == torch.Tensor: ``` This is incorrect because it does not handle subclasses of torch.Tensor. The correct...
## 📚 Documentation The "Pytorch/XLA overview" includes many advanced topics that go beyond an "overview", including how to specifically convert Stable Diffusion to run on TPUs (which is more of...
Docs build indicates that the docs have duplicate labels (aka anchors). These predate the recent changes to myst but now that we have standardized on the same tooling as upstream...
A) pypi (e.g. pyproject.toml) should present correct information to users B) longer-term, torch-xla should match torch supported python versions, which is listed in torch as >= 3.8.