Skip to content
Tiny TPU

Alternatives to Tiny TPU

200 apps like Tiny TPU, closest match first.

Top picks

  1. T

    tilegym-cutile-python

    Get

    Expert cuTile programming assistant. Write high-performance GPU kernels using cuTile's tile-based programming model with proper validation and optimization.

  2. T

    TrenTorch

    Launch

    TrenTorch is a set of free machine learning practice problems where you build the core ideas yourself. The main content is 1647 problems, each with a clear explanation, grouped into 12 topics and split into 466 beginner, 633 intermediate an

    Open source

  3. T

    tilegym-converting-cutile-to-triton

    Get

    Converts cuTile GPU kernels (@ct.kernel) to Triton (@triton.jit). Handles standard in-repo conversion, debugging (cudaErrorIllegalAddress, shape mismatch,

  4. R

    Rainbow Tensor

    Launch

    Colourful tensor operation visuals for NumPy style arrays, built for notebooks, docs, and teaching shape transformations.

    Open source

  5. T

    tilegym-improve-cutile-kernel-perf

    Get

    Iteratively optimize cuTile kernel performance through systematic profiling, bottleneck analysis, IR comparison, and targeted tuning. Covers tile sizes,

  6. M

    microgpt

    Launch

    microgpt is an interactive visualisation of a very small GPT model that runs in the browser. It is an educational tool for seeing how a language model generates text one character at a time. The model is trained on a dataset of names and le

More like Tiny TPU

7
H

HeyTensor

Free browser-based PyTorch tensor shape calculator with 22 tools

Launch
8
H

HappyTorch

A PyTorch coding practice platform — covering LLM, Diffusion, PEFT, and more…

Launch
9
T

Texts To Transformer

Train a tiny Transformer from scratch on your iMessage history

Launch
10
G

gke-ai-troubleshooting-tpu-vbar-oom

Diagnoses and prevents vbar_control_agent segfaults, out-of-memory

Get
11
T

tilegym-adding-cutile-kernel

Add a new cuTile GPU kernel operator to TileGym

Get
12
T

tilegym-cutile-autotuning

Adding, modifying, optimizing, or debugging CuTile autotuning code

Get
13
M

Mltu

Machine Learning Training Utilities (for TensorFlow and PyTorch)

Launch
14
D

docstring

Write docstrings for PyTorch functions and methods following PyTorch conventions

Get
15
S

SF Tensor

Infrastructure for AI labs to focus on research

Launch
16
T

Tiny Dec

A tiny educational decompiler that helps people understand how decompilation…

Launch
17
D

Deep Learning Studio

Design deep learning architectures, automate feature engineering

Launch
18
L

LPU Lite

A language processing unit built from first principles

Launch
19
T

turbo-operations

Turbo pipeline operations reference

Get
20
T

tilegym-converting-cutile-to-julia

Converts cuTile Python GPU kernels (@ct.kernel) to cuTile.jl Julia equivalents

Get
21
D

deep-learning-pytorch

Expert guidance for deep learning, transformers, diffusion models

Get
22
T

Time To First Token

A 10-week, 30-minutes-a-day roadmap for LLM inference serving and optimization

Launch
23
T

Tpot

A Python Automated Machine Learning tool

Launch
24
N

nemo-mbridge-perf-tp-dp-comm-overlap

Operational guide for enabling TP, DP

Get
25
N

nemo-mbridge-perf-moe-hardware-configs

Representative, point-in-time MoE training playbooks by hardware

Get
26
T

tao-setup-nvidia-gpu-host

Host setup for TAO GPU backends

Get
27
G

gke-ai-troubleshooting-tpu-dynamic-slices-monitoring

Monitors, troubleshoots, and manages GKE TPU Dynamic Slices custom resources

Get
28
A

academic-pipeline

Orchestrator for the full academic research pipeline: research -> write ->

Get
29
A

Atopile

Design circuit boards with code

Launch
30
N

nemo-mbridge-perf-moe-dispatcher-selection

Choose the right MoE token dispatcher

Get
31
T

Tutor GPT

AI tutor powered by Theory-of-Mind reasoning

Launch
32
T

Tulipcc

The Tulip Creative Computer and AMYboard - portable Python synthesizers

Launch
33
D

Docs

Documentations for PaddlePaddle

Launch
34
H

HipScript

Online compiler for HIP and NVIDIA CUDA code to WebGPU

Launch
35
T

tao-run-platform

TAO Execution SDK

Get
36
D

developer-tutorial

Writes or audits a teaching tutorial for a developer product

Get
37
G

gke-ai-troubleshooting-tpu-metrics-monitoring

Monitors and troubleshoots GKE TPU workloads, nodes, and node pools

Get
38
L

learn-mcp

Focused interactive tutor for the Model Context Protocol

Get
39
L

learn

Interactive lesson tutor for the AI Engineering from Scratch curriculum

Get
40
C

ChatTutor

Visual and Interactive AI Tutor

Launch
41
A

AI Infra Junior Engineer Learning

AI Infrastructure Junior Engineer Learning Track

Launch
42
S

SebastianGilPinzon/colab-mcp

Control Google Colab notebooks and assign GPUs (T4/L4/A100)

Get
43
S

Simten

Design hardware. Bit by bit

Launch
44
T

tao-run-automl-deft-pipeline

Run the canonical NVIDIA AOI three-phase training pipeline

Get
45
T

TensorTonic

Implement 200+ ML algorithms from scratch

Launch
46
H

huggingface-zerogpu

AI demos and GPU compute with Gradio Spaces and Hugging Face Spaces ZeroGPU

Get
47
T

tao-run-automl

Run container-backed AutoML / hyperparameter optimization

Get
48
A

algorithm-design

Design algorithms with LaTeX pseudocode and UML diagrams

Get
49
P

paper-assembly

Orchestrate the full paper pipeline end-to-end

Get
50
T

Texel.ai

Run AI pipelines 10x faster

Launch

About Tiny TPU

T

Tiny TPU

An attempt to understand and build a TPU—by complete novices

Tiny TPU is a write-up of an attempt by four complete novices to understand and build a tensor processing unit, the kind of accelerator Google uses for machine learning. The project is documented as a slideshow with sections on why the team started, what a TPU is, how they developed their toy design, and why they used XOR as a first problem. It walks through batching and dimensions, the systolic array of processing elements that performs matrix multiplication, input and weight scheduling with staggering and FIFOs, bias and leaky ReLU activation, pipelining and double buffering, and a forward pass walkthrough. The authors describe a strict philosophy of trying the hacky way first and consulting outside sources only after, and they say they wrote the code without relying on AI tools.

Price
Open source
Category
Design
Website
tinytpu.com