CodeZero

This release bumps the rl-swarm to run the CodeZero environment. For a full description of changes, please see our documentation: https://docs.gensyn.ai/testnet/rl-swarm/how-it-works/codezero

v0.6.4

Full Changelog: v0.6.3...v0.6.4

• Catching more errors to protect from malicious payloads
• Bumping GenRL tag

v0.6.3

Full Changelog: v0.6.2...v0.6.3

• Adding try-except block in GenRL for running the game round. If there are exceptions caught, continue the round with just your own payload
• Removed logging statement from DHT

v0.6.2

Attempt to filter bad actors with stringent type checking. There are many nodes submitting ill-formed payloads into the DHT, this attempts to prevent those bad actors from crashing the DHT.

v0.6.1

Adds error handling around DHT deserialization. This update is only needed if your node has been crashing when performing the all_gather from the DHT. Some objects were being serialized that were not supported by the nodes, this will skip those objects.

v0.6.0

Full Changelog: v0.5.8...v0.6.0

Nightly: v0.5.8

Nightly Release: Includes

• patch for max() iterable empty error.
• Debug info and instructions for nodes that receive HTTP 400 Error on submit rewards. If encountered, please file a github issue and include the full stacktrace in the issue body.

v0.5.7

submit TXNs to chain based on time elapsed, not rounds completed

* submit based on time, not rounds

* submit at end of round if missed opportunity during rewards calc

v0.5.6

Hotfix: reduce pressure on testnet to prevent network outages.

v0.5.5

Small fix to catch the case when a node does not communicate an action