Home/Model Training/octo

octo

octo-models/octo

Transformer-based robot policy trained on a diverse mix of robot trajectories.

GitHub Website

GraphCanon updated 3w · GitHub synced 3w

1.7k stars276 forksLast push 2y Python MIT

Decision brief

Octo focuses on transformer-based models for robot control, emphasizing diverse trajectory training and compatibility with both GPU and TPU via Jax.

Good fit when

Need advanced model finetuning with a pre-existing transformer foundation
Working on hardware accelerated applications leveraging GPUs or TPUs

Avoid when

If your project requires real-time decision-making without access to GPU/TPU resources
Looking for simpler, more generalized model training tools outside robot control tasks

Observed Jul 12, 2026 · Source: enrich:decision_facts

octo alternatives Trust report

Verify the decision

Maintenance and security

Full trust report

Maintenance: Dormant (731d since push); As of 3w
Provenance: Not a fork · Organization account; As of 3w
Security (OSV): 48 low (48 low); As of 1mo

Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.

Install

pip install octo

PyPI

Similar tools

Same-category neighbours. No typed graph edges are catalogued for this tool yet.

tensorflow

An Open Source Machine Learning Framework for Everyone

197k★

transformers

Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models

164k★

langflow

Tool for building and deploying AI-powered agents and workflows

153k★

Evidence and technical details

Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.

Overview

Octo includes code for training and finetuning transformer-based models aimed at controlling robots, leveraging Jax for GPU and TPU support. The repository emphasizes installation setup for different hardware configurations to facilitate model training.

Capability facts

Languages: python
Source: github.language+pyproject.toml · Aug 2, 2026

Compatibility

Sourced claims from the README excerpt - not unsourced marketing copy.

Python runtimePython

Source: README excerpt (regex_v1, Aug 2, 2026)

conda create -n octo python=3.10

Source link

README

Installation

conda create -n octo python=3.10
conda activate octo
pip install -e .
pip install -r requirements.txt

For GPU:

pip install --upgrade "jax[cuda11_pip]==0.4.20" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html

For TPU

pip install --upgrade "jax[tpu]==0.4.20" -f https://storage.googleapis.com/jax-releases/libtpu_releases.html

See the Jax Github page for more details on installing Jax.

Test the installation by finetuning on the debug dataset:

python scripts/finetune.py --config.pretrained_path=hf://rail-berkeley/octo-small-1.5 --debug

For agents

This page has a .md twin and JSON over the API.

/tools/octo-models-octo.md GET /api/graphcanon/tools/octo-models-octo GET /api/graphcanon/graph?tool=octo-models-octo