AI Hypercomputer

All

20 repositories

maxtext
Public
A simple, performant and scalable Jax LLM!
gpt mistral fine-tuning jax sft large-language-models llm llama2 mixtral deepseek
Python
•
Apache License 2.0
•415•1.9k•60•266•Updated Sep 13, 2025Sep 13, 2025
RecML
Public
Python
•
Apache License 2.0
•17•208•3•33•Updated Sep 13, 2025Sep 13, 2025
maxdiffusion
Public
Python
•
Apache License 2.0
•41•258•4•13•Updated Sep 13, 2025Sep 13, 2025
gpu-recipes
Public
Recipes for reproducing training and serving benchmarks for large machine learning models using GPUs on Google Cloud.
gpu benchmarks google-cloud-platform serving distributed-training
Python
•
Apache License 2.0
•33•83•4•6•Updated Sep 12, 2025Sep 12, 2025
torchprime
Public
torchprime is a reference model implementation for PyTorch on TPU.
Python
•
Other
•9•35•89•13•Updated Sep 12, 2025Sep 12, 2025
xpk
Public
xpk (Accelerated Processing Kit, pronounced x-p-k,) is a software tool to help Cloud developers to orchestrate training jobs on accelerators such as TPUs and GPUs on GKE.
gke gcloud tpu
Python
•
Apache License 2.0
•49•141•24•39•Updated Sep 12, 2025Sep 12, 2025
ml-goodput-measurement
Public
Python
•
Apache License 2.0
•3•21•1•0•Updated Sep 11, 2025Sep 11, 2025
pathways-utils
Public
Package of Pathways-on-Cloud utilities
Python
•
Apache License 2.0
•4•17•0•3•Updated Sep 4, 2025Sep 4, 2025
resiliency
Public
This project provides resilient training logic to improve the training efficiency of large-scale distributed workloads on Google Cloud.
Python
•
Apache License 2.0
•0•6•0•0•Updated Sep 4, 2025Sep 4, 2025
accelerator-microbenchmarks
Public
Python
•
Apache License 2.0
•7•3•0•3•Updated Sep 2, 2025Sep 2, 2025
cloud-accelerator-diagnostics
Public
Python
•
Apache License 2.0
•10•23•4•0•Updated Aug 31, 2025Aug 31, 2025
cloud-diagnostics-xprof
Public
Python
•
Apache License 2.0
•1•11•0•0•Updated Aug 20, 2025Aug 20, 2025
tpu-recipes
Public
Shell
•
Apache License 2.0
•19•45•3•7•Updated Aug 19, 2025Aug 19, 2025
inference-benchmark
Public
Python
•
Apache License 2.0
•13•16•7•4•Updated Jun 18, 2025Jun 18, 2025
JetStream
Public
JetStream is a throughput and memory optimized engine for LLM inference on XLA devices, starting with TPUs (and GPUs in future -- PRs welcome).
gpu inference pytorch transformer llama gpt gemma model-serving tpu jax
Python
•
Apache License 2.0
•51•375•14•12•Updated Jun 10, 2025Jun 10, 2025
kithara
Public
Python
•
Apache License 2.0
•6•14•0•2•Updated May 19, 2025May 19, 2025
ray-tpu
Public
Python
•
Apache License 2.0
•4•15•3•0•Updated May 11, 2025May 11, 2025
aotc
Public
Python
•
Other
•1•0•1•0•Updated May 8, 2025May 8, 2025
jetstream-pytorch
Public
PyTorch/XLA integration with JetStream (https://github.com/google/JetStream) for LLM inference"
inference pytorch batching attention llama gemma model-serving tpu llm llm-inference
Python
•
Apache License 2.0
•18•71•10•4•Updated Mar 27, 2025Mar 27, 2025
cloud-tpu-monitoring-debugging
Public
HCL
•
Apache License 2.0
•1•13•0•0•Updated Feb 18, 2025Feb 18, 2025