Sergii Dymchenko
Sergii Dymchenko
torch.qr is deprecated for a long time and is being removed by https://github.com/pytorch/pytorch/pull/70989. This PR makes the example compatible with new and old PyTorch versions.
PyTorch's `torch.eig` was deprecated since version 1.9 and is being removed by https://github.com/pytorch/pytorch/pull/70982. Please use the `torch.linalg.eig` function instead if you want your code to continue to work with the...
Fix require_grad typos (should be requires_grad). Before the fix, the code doesn't cause any errors but doesn't do what it's supposed to do. Also see https://github.com/pytorch/benchmark/pull/1771
For example, I want to replace all args of a `Call`: ``` node.with_deep_changes( old_node=cst.ensure_type(node.value, cst.Call).args, value=[input_arg] ) ``` The code works fine, but the type checker complains: error: Argument "old_node"...
Here is a full repro: ``` python import libcst as cst def f(node: cst.Call) -> cst.Call: arg = node.args[0] new_arg = arg.with_changes(value=cst.Integer(value="2")) return node.deep_replace(arg, new_arg) node = cst.ensure_type(cst.parse_expression("func(1)"), cst.Call) print(f(node))...
Not sure if this should be considered incorrect, but it was surprising for me. With code like this `vmap(a)(b)`, and the tree like this: ``` Call( func=Call( func=Name( value='vmap', ```...
I'm using `parallel_exec_transform_with_prettyprint` from my own command-line script and it prints a lot of "Failed to determine module name" messages from https://github.com/Instagram/LibCST/blob/main/libcst/codemod/_cli.py#L265 Is there a good way to silence that?
This is found via https://github.com/pytorch-labs/torchfix/ `torch.load` without `weights_only` parameter is unsafe. Explicitly set `weights_only` to False only if you trust the data you load and full pickle functionality is needed,...
## Description `torch.load` without `weights_only` parameter is unsafe. Explicitly set `weights_only` to False only if you trust the data you load and full pickle functionality is needed, otherwise set `weights_only=True`....