Skip to main content

Reverb is an efficient and easy-to-use data storage and transport system designed for machine learning research.

Project description


PyPI - Python Version PyPI version

Reverb is an efficient and easy-to-use data storage and transport system designed for machine learning research. Reverb is primarily used as an experience replay system for distributed reinforcement learning algorithms but the system also supports multiple data structure representations such as FIFO, LIFO, and priority queues.

Table of Contents


Please keep in mind that Reverb is not hardened for production use, and while we do our best to keep things in working order, things may break or segfault.

:warning: Reverb currently only supports Linux based OSes.

The recommended way to install Reverb is with pip. We also provide instructions to build from source using the same docker images we use for releases.

TensorFlow can be installed separately or as part of the pip install. Installing TensorFlow as part of the install ensures compatibility.

$ pip install dm-reverb[tensorflow]

# Without Tensorflow install and version dependency check.
$ pip install dm-reverb

Nightly builds

PyPI version

$ pip install dm-reverb-nightly[tensorflow]

# Without Tensorflow install and version dependency check.
$ pip install dm-reverb-nightly

Build from source

This guide details how to build Reverb from source.

Quick Start

Starting a Reverb server is as simple as:

import reverb

server = reverb.Server(tables=[

Create a client to communicate with the server:

client = reverb.Client(localhost:8000)

Write some data to the table:

# Creates a single item and data element [0, 1].
client.insert([0, 1], priorities={'my_table': 1.0})

An item can also reference multiple data elements:

# Creates three data elements ([2, 2] , [3, 3], and [4, 4]) and a single item
# `[[2, 2], [3, 3], [4, 4]]` that references all three of them.
with client.writer(max_sequence_length=3) as writer:
  writer.append([2, 2])
  writer.append([3, 3])
  writer.append([4, 4])
  writer.create_item('my_table', num_timesteps=3, priority=1.0)

The items we have added to Reverb can be read by sampling them:

# client.sample() returns a generator.
print(list(client.sample('my_table', num_samples=2)))

Continue with the Reverb Tutorial for an interactive tutorial.

Detailed overview

Experience replay has become an important tool for training off-policy reinforcement learning policies. It is used by algorithms such as Deep Q-Networks (DQN), Soft Actor-Critic (SAC), Deep Deterministic Policy Gradients (DDPG), and Hindsight Experience Replay, ... However building an efficient, easy to use, and scalable replay system can be challenging. For good performance Reverb is implemented in C++ and to enable distributed usage it provides a gRPC service for adding, sampling, and updating the contents of the tables. Python clients expose the full functionality of the service in an easy to use fashion. Furthermore native TensorFlow ops are available for performant integration with TensorFlow and

Although originally designed for off-policy reinforcement learning, Reverb's flexibility makes it just as useful for on-policy reinforcement -- or even (un)supervised learning. Creative users have even used Reverb to store and distribute frequently updated data (such as model weights), acting as an in-memory light-weight alternative to a distributed file system where each table represents a file.


A Reverb Server consists of one or more tables. A table hold items, and each item references one or more data elements. Tables also define sample and removal selection strategies, a maximum item capacity, and a rate limiter.

Multiple items can reference the same data element, even if these items exist in different tables. This is because items only contain references to data elements (as opposed to a copy of the data itself). This also means that a data element is only removed when there exists no item that contains a reference to it.

For example, it is possible to set up one Table as a Prioritized Experience Replay (PER) for transitions (sequences of length 2), and another Table as a (FIFO) queue of sequences of length 3. In this case the PER data could be used to train DQN, and the FIFO data to train a transition model for the environment.

Using multiple tables

Items are automatically removed from the Table when one of two conditions are met:

  1. Inserting a new item would cause the number of items in the Table to exceed its maximum capacity.

  2. An item has been sampled more than the maximum number of times permitted by the Table’s rate limiter. Note that not all rate limiters will enforce this.

In both cases, which item to remove is determined by the table’s removal strategy. As mentioned earlier, a data element is automatically removed from the Server when the number of items that references it reaches zero.

Users have full control over how data is sampled and removed from Reverb tables. The behavior is primarily controlled by the item selection strategies provided to the Table as the sampler and remover. In combination with the rate_limiter and max_times_sampled, a wide range of behaviors can be achieved. Some commonly used configurations include:

Uniform Experience Replay

A set of the N=1000 most recently inserted items are maintained. By setting sampler=reverb.selectors.Uniform(), the probability to select an item is the same for all items. Due to reverb.rate_limiters.MinSize(100), sampling requests will block until 100 items have been inserted. By setting remover=reverb.selectors.Fifo() when an item needs to be removed the oldest item is removed first.


Examples of algorithms that make use of uniform experience replay include SAC and DDPG.

Prioritized Experience Replay

A set of the N=1000 most recently inserted items. By setting sampler=reverb.selectors.Prioritized(priority_exponent=0.8), the probability to select an item is proportional to the item's priority.

Note: See Schaul, Tom, et al. for the algorithm used in this implementation of Prioritized Experience Replay.


Examples of algorithms that make use of Prioritized Experience Replay are DQN (and its variants), and Distributed Distributional Deterministic Policy Gradients.


Collection of up to N=1000 items where the oldest item is selected and removed in the same operation. If the collection contains 1000 items then insert calls are blocked until it is no longer full, if the collection is empty then sample calls are blocked until there is at least one item.


# Or use the helper classmethod `.queue`.
reverb.Table.queue(name=my_queue', max_size=1000)

Examples of algorithms that make use of Queues are IMPALA and asynchronous implementations of Proximal Policy Optimization.

Item selection strategies

Reverb defines several selectors that can be used for item sampling or removal:

  • Uniform: Sample uniformly among all items.
  • Prioritized: Samples proportional to stored priorities.
  • FIFO: Selects the oldest data.
  • LIFO: Selects the newest data.
  • MinHeap: Selects data with the lowest priority.
  • MaxHeap: Selects data with the highest priority.

Any of these strategies can be used for sampling or removing items from a Table. This gives users the flexibility to create customized Tables that best fit their needs.

Rate Limiting

Rate limiters allow users to enforce conditions on when items can be inserted and/or sampled from a Table. Here is a list of the rate limiters that are currently available in Reverb:

  • MinSize: Sets a minimum number of items that must be in the Table before anything can be sampled.
  • SampleToInsertRatio: Sets that the average ratio of inserts to samples by blocking insert and/or sample requests. This is useful for controlling the number of times each item is sampled before being removed.
  • Queue: Items are sampled exactly once before being removed.
  • Stack: Items are sampled exactly once before being removed.


Reverb servers are unaware of each other and when scaling up a system to a multi server setup data is not replicated across more than one node. This makes Reverb unsuitable as a traditional database but has the benefit of making it trivial to scale up systems where some level of data loss is acceptable.

Distributed systems can be horizontally scaled by simply increasing the number of Reverb servers. When used in combination with a gRPC compatible load balancer, the address of the load balanced target can simply be provided to a Reverb client and operations will automatically be distributed across the different nodes. You'll find details about the specific behaviors in the documentation of the relevant methods and classes.

If a load balancer is not available in your setup or if more control is required then systems can still be scaled in almost the same way. Simply increase the number of Reverb servers and create separate clients for each server.


Reverb supports checkpointing; the state and content of Reverb servers can be stored to permanent storage. While pointing, the Server serializes all of its data and metadata needed to reconstruct it. During this process the Server blocks all incoming insert, sample, update, and delete requests.

Checkpointing is done with a call from the Reverb Client:

# client.checkpoint() returns the path the checkpoint was written to.
checkpoint_path = client.checkpoint()

To restore the a reverb.Server from a checkpoint:

checkpointer = reverb.checkpointers.DefaultCheckpointer(path=checkpoint_path)
# The arguments passed to `tables=` must be the as those used by the `Server`
# that wrote the checkpoint.
server = reverb.Server(tables=[...], checkpointer=checkpointer)

Refer to tfrecord_checkpointer.h for details on the implementation of checkpointing in Reverb.


If you use this code, please cite it as:

  title = {{Reverb}: An efficient data storage and transport system for ML research},
  author = "{Albin Cassirer, Gabriel Barth-Maron, Thibault Sottiaux, Manuel Kroiss, Eugene Brevdo}",
  howpublished = {\url{}},
  url = "",
  year = 2020,
  note = "[Online; accessed 01-June-2020]"

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for dm-reverb, version 0.2.0
Filename, size File type Python version Upload date Hashes
Filename, size dm_reverb-0.2.0-cp36-cp36m-manylinux2010_x86_64.whl (6.3 MB) File type Wheel Python version cp36 Upload date Hashes View
Filename, size dm_reverb-0.2.0-cp37-cp37m-manylinux2010_x86_64.whl (6.3 MB) File type Wheel Python version cp37 Upload date Hashes View
Filename, size dm_reverb-0.2.0-cp38-cp38-manylinux2010_x86_64.whl (6.3 MB) File type Wheel Python version cp38 Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page