Yuchao Zhang

any updates here?

According to Python doc of `debug` for example, this function (as well as [info()](https://docs.python.org/3/library/logging.html#logging.info), [warning()](https://docs.python.org/3/library/logging.html#logging.warning), [error()](https://docs.python.org/3/library/logging.html#logging.error) and [critical()](https://docs.python.org/3/library/logging.html#logging.critical)) will call [basicConfig()](https://docs.python.org/3/library/logging.html#logging.basicConfig) if the root logger doesn’t have any handler attached....

same issue here. Don't know if torch.autocast can be used together with deepspeed fp16

Can be leverage this from C++ community? https://github.com/kirkshoop/async_scope/blob/main/asyncscope.md

Same issue here, even in azureml public docker. ``` Collecting environment information... PyTorch version: 2.2.1 Is debug build: False CUDA used to build PyTorch: 12.1 ROCM used to build PyTorch:...

What about the async scope (scoped spawn) design for C++'s p2300 sender/receiver async proposal? https://github.com/kirkshoop/async_scope/blob/main/asyncscope.md As the paper says, since not every async-function has a clear chain of work to...

I don't quite get the context. What do you mean we can't depend on Drop?

> The answer is that crossbeam does not rely on `Drop` for this. They use a `scope` function, and if any of the threads are still running when you return...

@Darksonn you mentioned [here](https://www.reddit.com/r/rust/comments/wltcf8/comment/ijwx3mz/) it's related to cancelation. Can u explain that? Thanks.

@Darksonn What about add one layer of abstraction like [async resource](https://github.com/kirkshoop/async_scope/blob/main/async-resource.md) to make sure attached resources like `data` will only be destroyed after all tasks complete. ``` rt::block_on(use_resource( |scope, data|...