# google-deepmind/scalable_agent

A TensorFlow implementation of Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures.

Repository: https://github.com/google-deepmind/scalable_agent
Canonical: https://ross.abutalabs.com/products/scalable_agent
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2019-03-13T04:43:25+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3010, "days_push": 2730, "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 1028, forks 161 (observed 2026-08-28T04:03:17.531444+00:00)

## What it is
A TensorFlow implementation of IMPALA (Importance Weighted Actor-Learner Architectures) for scalable distributed deep reinforcement learning, including a dynamic batching module. It supports single-machine and distributed training on DeepMind Lab environments such as DMLab-30.

## Use cases
- train deep RL agents at scale with distributed actor-learner architecture
- reproduce IMPALA results on DMLab-30
- run reinforcement learning experiments on DeepMind Lab levels
- use importance-weighted V-trace off-policy correction
- dynamically batch experience from many parallel actors

## When to choose
- you want a reference implementation of the IMPALA algorithm for research
- you need scalable distributed deep RL with many parallel environment actors
- you work with DeepMind Lab or similar environments in TensorFlow

## When to avoid
- you need a maintained production library - the repo has not been updated since 2019
- you prefer PyTorch or modern RL frameworks
- you need official Google support or long-term compatibility with recent TensorFlow versions

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, reinforcement-learning, deep-learning, streaming
- domain: reinforcement-learning, machine-learning, deep-learning
- platform: python
- tags: impala, actor-learner, tensorflow, deepmind-lab, distributed-training, research-code, linux, docker

## Member repositories
- google-deepmind/scalable_agent (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:17.531444+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 2026-08-30T07:07:45.673871+00:00, confidence not recorded.
  - readme: https://github.com/google-deepmind/scalable_agent (fetched 2026-08-28T04:03:17.531444+00:00, sha 2e858172cf4f)
- Data as of 2026-08-30T08:39:29.467469+00:00.