Alex Morehead

This issue seems to be happening for me with AMD MI250xs and `wandb==0.17.8`, while using WandB's new `core` backend. Sadly, until this issue is patched, I'll have to revert to...

@dmitryduev, thank you very much for your quick patch for this issue! I can confirm on my end that the ROCm (MI250x) system I am using is now properly having...

@dmitryduev, to confirm, is this fix included in `wandb==0.18.0`? I ask because I didn't see it listed in 0.18.0's release notes. I suspect the answer is yes per https://github.com/wandb/wandb/commits/v0.18.0/, but...

@dmitryduev @kptkin, I noticed today that with `torch>2.2` and ROCm 6.1/6.2 (e.g., torch 2.3.0 or 2.4.1), `wandb==0.18.0` captures all relevant ROCm GPU metrics except for GPU memory usage (e.g., GPU...

> Yes this is expected. Lightning can't know how to shard the data/iterator you provide. You need to make sure your iterator returns half of the data on GPU 0...

@awaelchli, I second this issue. I am also having difficulties figuring out the simplest way to enable multi-GPU and multi-dataloader worker support for IterableDatasets when using PyTorch Lightning. All the...

I believe [this PR](https://github.com/zrqiao/NeuralPLexer/pull/27) addresses the missing plDDT values issue.

Hi, @wanggaa. Thanks for your kind words on this project! My intuition tells me that this issue is caused by said ions having a zero vector for their coordinates, as...

Hi, @zqcai19. My first thoughts are that these differences may be the result of the PDB dataset's preprocessing scripts, as described in the AF3 paper. This preprocessing script will (in...

Hi, @zqcai19. This filtering of the train, val, and test structures (particularly for the test structures) seems to be implicitly suggested by the AF3 paper. To standardize all three dataset...