xla
gloo
xla | gloo | |
---|---|---|
8 | 2 | |
2,296 | 1,140 | |
1.4% | 0.7% | |
9.9 | 8.0 | |
1 day ago | 6 days ago | |
C++ | C++ | |
GNU General Public License v3.0 or later | GNU General Public License v3.0 or later |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
xla
-
Who uses Google TPUs for inference in production?
> The PyTorch/XLA Team at Google
Meanwhile you have an issue from 5 years ago with 0 support
https://github.com/pytorch/xla/issues/202
-
Google TPU v5p beats Nvidia H100
PyTorch has had an XLA backend for years. I don't know how performant it is though. https://pytorch.org/xla
-
Why Did Google Brain Exist?
It's curtains for XLA, to be precise. And PyTorch officially supports XLA backend nowadays too ([1]), which kind of makes JAX and PyTorch standing on the same foundation.
1. https://github.com/pytorch/xla
-
Accelerating AI inference?
Pytorch supports other kinds of accelerators (e.g. FPGA, and https://github.com/pytorch/glow), but unless you want to become a ML systems engineer and have money and time to throw away, or a business case to fund it, it is not worth it. In general, both pytorch and tensorflow have hardware abstractions that will compile down to device code. (XLA, https://github.com/pytorch/xla, https://github.com/pytorch/glow). TPUs and GPUs have very different strengths; so getting top performance requires a lot of manual optimizations. Considering the the cost of training LLM, it is time well spent.
-
[D] Colab TPU low performance
While apparently TPUs can theoretically achieve great speedups, getting to the point where they beat a single GPU requires a lot of fiddling around and debugging. A specific setup is required to make it work properly. E.g., here it says that to exploit TPUs you might need a better CPU to keep the TPU busy, than the one in colab. The tutorials I looked at oversimplified the whole matter, the same goes for pytorch-lightning which implies switching to TPU is as easy as changing a single parameter. Furthermore, none of the tutorials I saw (even after specifically searching for that) went into detail about why and how to set up a GCS bucket for data loading.
- How to train large deep learning models as a startup
-
Distributed Training Made Easy with PyTorch-Ignite
XLA on TPUs via pytorch/xla.
-
[P] PyTorch for TensorFlow Users - A Minimal Diff
I don't know of any such trick except for using TensorFlow. In fact, I benchmarked PyTorch XLA vs TensorFlow and found that the former's performance was quite abysmal: PyTorch XLA is very slow on Google Colab. The developers' explanation, as I understood it, was that TF was using features not available to the PyTorch XLA developers and that they therefore could not compete on performance. The situation may be different today, I don't know really.
gloo
-
Releasing Gloo 0.4.0
These are two separate libraries that do very different things but share the same name. They are also written in two separate languages. That is a sizable gap between them, and reusing names happens often with libraries. Gloo (rust-wasm, this post) is also not new. Though, relative to Gloo (Go, solo-io), it is newer. But, there is also a Github repo even older than Gloo (solo-io): https://github.com/facebookincubator/gloo. As well, even if these were for some odd reason all about wasm, none of them are actually that popular. solo-io Gloo has the most stars (though that isn't the best metric of popularity, since it is relative to the community that actually uses it), but 3k simply isn't that much. There is certainly a good argument to look down on libraries that reuse popular library names, but this isn't really the case here. Both started not too long after each other (solo-io would not have most of the stars it currently has when Gloo-Rust started), are in separate languages (thus separate communities), and do very separate things.
-
Distributed Training Made Easy with PyTorch-Ignite
backends from native torch distributed configuration: nccl, gloo, mpi.
What are some alternatives?
NCCL - Optimized primitives for collective multi-GPU communication
pytorch-lightning - Build high-performance AI models with PyTorch Lightning (organized PyTorch). Deploy models with Lightning Apps (organized Python to build end-to-end ML systems). [Moved to: https://github.com/Lightning-AI/lightning]
ompi - Open MPI main development repository
why-ignite - Why should we use PyTorch-Ignite ?
gloo - The Feature-rich, Kubernetes-native, Next-Generation API Gateway Built on Envoy
pocketsphinx - A small speech recognizer
ignite - High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.
idist-snippets