Zola
Zola
Hi @cpuhrsch, To my understanding, as part of the tests using `torch/testing/_internal/common_methods_invocations.py`'s `OpInfo`s, there are parametrized tests testing the ops with all supported dtypes as well as supposedly unsupported dtypes....
@ZainRizvi the test fail is caused by #128637, not this PR ([as seen here](https://github.com/pytorch/pytorch/actions/runs/10604690693/job/29393478235#step:20:3075))
Hey! Any news on this one?
Also, would it be possible not to warn when only `MASTER_ADDR` and `MASTER_PORT` are provided as there are used and seem to be intended to be provided https://github.com/pytorch/ignite/blob/34a707e53785cf8a524589f33a570a7516fe064e/ignite/distributed/comp_models/native.py#L607-L614
Hi, > There is no env var responsible for that this argument ? By the way, why this is necessary to set it and what's the typical value, 1 ?...
Hi @vfdev-5, > Following https://slurm.schedmd.com/sbatch.html, SLURM_NTASKS_PER_GPU is set when --ntasks-per-gpu is specified. `SLURM_NTASKS_PER_GPU` is just set by `sbatch` as an output variable so that it can be used as input...
Seems related to #4353
Hi, Yes, the `--no-lock` flag seems to fix the issue! However @travishathaway I don't have a lot of information regarding the file system, I can ask to the sysadmin. It...
Hi @vfdev-5, Yes exactly, the code snippet you provided is a good example. Another example would be the following: ```python evaluator.state.metrics = ... # kept unchanged handler = OutputHandler( tag="validation",...
Before testing this, I would like to be sure the feature is of interest.