-
- [x] Learn the basics
- [x] Write simple hello
- [x] Write a pipeline code using transformer.
-
# Steps to Reproduce
Go to any populated sim ~ pull up ARC and read.
I teleported around to various busy sims and looked at the general population, collecting info about average ARC / Render Weig…
-
We know that flash attention supports `cu_seqlens`, which can remove padding for variable-length input in a batch and only store regular tokens. This can be useful for optimizing the computational eff…
-
### Your current environment
Collecting environment information...
PyTorch version: 2.3.1+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A
OS: Ubu…
-
### Prerequisite
- [X] I have searched [Issues](https://github.com/open-mmlab/mmdetection3d/issues) and [Discussions](https://github.com/open-mmlab/mmdetection3d/discussions) but cannot get the expec…
-
In the attached drafts I have tried to incorporate the changes relative to the 20170708 versions that appeared to be consensus (at least more than one supporter).
The biggest change is a retro-grad…
-
RL is flaky and one needs to report statistically significant results. We should help contributors and users with this. ~~I think using nni for parallel execution is a good fit for that~~ (nni is dead…
-
Dear s2fft devs,
Thanks for this great package! I just wanted to report on some of the pain-points I ran into to hopefully provide constructive feedback.
#### SHT with inappropriate healpix nsid…
-
**Problem**
The Client-Server API is not bandwidth efficient enough. matrix-org/matrix-spec-proposals#3079 was supposed to help with that, but it has too broad scope and has the own issues. One of th…
-
Supercycling particle integration implemented in https://github.com/MPAS-Dev/MPAS/pull/1484 should be verified for global runs.