Topics / #tpu

Tpu projects on GitHub

Training and running machine learning workloads on Google Tensor Processing Units.

As of , the most-starred Tpu projects on GitHub are vllm (94.1K stars), sglang (37.3K stars) and jax (36.5K stars). GitHub.Locker tracks 8 Tpu repositories with 223.1K stars combined. The fastest-growing on October 17 was vllm, with +69 stars in a day.

8Projects
223.1KStars in total
+127Stars in a day
Browse in the catalog →github.com/topics/tpu ↗

Hot right now

Most stars gained in a day · Oct 17

+69

vllm

vllm-project/vllm · ★ 94.1K

vLLM is a library for fast LLM inference and serving, built around PagedAttention for efficient memory management of attention keys and values.

+30

sglang

sgl-project/sglang · ★ 37.3K

SGLang is a high-performance serving framework for large language models and multimodal models, built to deliver low-latency, high-throughput inference from a single GPU up to large distributed clusters.

+14

kaggle-tpu-lab

ARahim3/kaggle-tpu-lab · ★ 670

A recipe that serves frontier-class open LLMs — GLM-5.3-Flash 320B MoE and Qwen3.8-27B bf16 — on Kaggle's free TPU v5e-8, exposing a public OpenAI-compatible (and Anthropic-compatible) endpoint you can plug into coding agents like Claude Code, Codex, opencode, or pi.

+8

jax

jax-ml/jax · ★ 36.5K

JAX is a Python library for accelerator-oriented array computation and program transformation, built for high-performance numerical computing and large-scale machine learning.

The all-time top

8 projects · by stars

  1. #1 · vllm-project/vllm

    vllm

    vLLM is a library for fast LLM inference and serving, built around PagedAttention for efficient memory management of attention keys and values.

    94.1K
  2. #2 · sgl-project/sglang

    sglang

    SGLang is a high-performance serving framework for large language models and multimodal models, built to deliver low-latency, high-throughput inference from a single GPU up to large distributed clusters.

    37.3K
  3. #3 · jax-ml/jax

    jax

    JAX is a Python library for accelerator-oriented array computation and program transformation, built for high-performance numerical computing and large-scale machine learning.

    36.5K
  1. 4

    tensor2tensor

    17.5Ktensorflow/tensor2tensor · Python

    A library of deep learning models and datasets from researchers and engineers on the Google Brain team, built to make deep learning more accessible and speed up ML research.

  2. 5

    litgpt

    13.7KLightning-AI/litgpt · Python

    LitGPT is a collection of 20+ high-performance large language models (Llama, Gemma, Phi, Qwen, Mistral, Falcon, TinyLlama and more) implemented from scratch, with ready-made recipes to pretrain, finetune, evaluate and deploy them at scale.

  3. 6

    vision_transformer

    12.7Kgoogle-research/vision_transformer · Jupyter Notebook

    This is Google Research's official JAX/Flax code release for Vision Transformer (ViT) and MLP-Mixer image classification models, including pretrained checkpoints and scripts to fine-tune them on your own data.

  4. 7

    skypilot

    10.7Kskypilot-org/skypilot · Python

    SkyPilot is an open-source system for running, managing, and scaling AI workloads — training, serving, batch inference — on any infrastructure: your own Kubernetes or Slurm clusters, reserved GPUs, or 20+ clouds.

  5. 8

    kaggle-tpu-lab

    670ARahim3/kaggle-tpu-lab · Python

    A recipe that serves frontier-class open LLMs — GLM-5.3-Flash 320B MoE and Qwen3.8-27B bf16 — on Kaggle's free TPU v5e-8, exposing a public OpenAI-compatible (and Anthropic-compatible) endpoint you can plug into coding agents like Claude Code, Codex, opencode, or pi.

Often tagged together