# tensorflow/agents

TF-Agents: A reliable, scalable and easy to use TensorFlow library for Contextual Bandits and Reinforcement Learning.

Repository: https://github.com/tensorflow/agents
Canonical: https://ross.abutalabs.com/products/tensorflow-agents
Language: Python
License: Apache-2.0
License Family: permissive
Topics: reinforcement-learning, tensorflow, contextual-bandits, bandits, multi-armed-bandits, tf-agents, rl-algorithms, dqn
Last push: 2026-01-16T22:49:15+00:00

## Health v2 (maintenance only)
Score: 60/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 62, release rhythm 35, longevity 100
- inputs: {"age_days": 2847, "days_push": 229, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3025, forks 752 (observed 2026-08-28T04:07:38.638981+00:00)

## What it is
TF-Agents is a TensorFlow-based Python library for building, training, and deploying reinforcement learning and contextual bandit algorithms. It provides well-tested modular components such as agents, policies, and replay buffers, along with implementations of algorithms like DQN, DDPG, TD3, PPO, SAC, and REINFORCE.

## Use cases
- train a DQN agent on Cartpole
- implement a custom reinforcement learning algorithm in TensorFlow
- solve contextual bandit problems
- prototype and benchmark RL algorithms
- train a soft actor-critic agent for continuous control
- learn reinforcement learning through tutorials

## When to choose
- you want reliable, modular, well-tested RL components in TensorFlow
- you need standard algorithms like DQN, PPO, or SAC out of the box
- you are working on contextual bandits or multi-armed bandits
- you want fast iteration with good test integration and benchmarking

## When to avoid
- you prefer PyTorch over TensorFlow
- you need a fully stable API with no breaking changes
- you only need lightweight RL without deep learning frameworks

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, reinforcement-learning, benchmarking, sdk
- domain: reinforcement-learning, machine-learning, deep-learning
- platform: python
- tags: tensorflow, reinforcement-learning, contextual-bandits, multi-armed-bandits, dqn, ppo, sac, rl-algorithms

## Member repositories
- tensorflow/agents (main) score 60

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:38.638981+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:29:54.664036+00:00, confidence not recorded.
  - readme: https://github.com/tensorflow/agents (fetched 2026-08-28T04:07:38.638981+00:00, sha 50d60fb439d3)
  - registry_pypi: https://pypi.org/pypi/agents/json (fetched 2026-08-29T09:44:55.423464+00:00, sha ed4f763c6319)
- Data as of 2026-08-30T08:39:29.467469+00:00.