v0.5.0

ReinforcementLearningCore v0.5.0

Diff since v0.4.6

Merged pull requests:

• Improve tabular learner (#140) (@findmyway)
• minor fix (#142) (@findmyway)
• Split the logic of run into two steps to allow injecting checkers (#143) (@findmyway)
• Replace reservoir array with reservoir trajectory (#144) (@findmyway)

v0.4.6

ReinforcementLearningCore v0.4.6

Diff since v0.4.5

Merged pull requests:

• CompatHelper: bump compat for "Distributions" to "0.24" (#139) (@github-actions[bot])

v0.4.5

ReinforcementLearningCore v0.4.5

Diff since v0.4.4

Closed issues:

• Performance issue with models running on CPU (#130)

Merged pull requests:

• updated send_to_device, for ElasticArray type. (#127) (@norci)
• minor enhancement (#128) (@findmyway)
• updated send_to_device for ElasticArray. (#131) (@norci)
• CompatHelper: bump compat for "GPUArrays" to "6.0" (#132) (@github-actions[bot])
• Add reservoir_array_buffer (#133) (@findmyway)
• normlogpdf - workaround for GPU differentiability (#134) (@rbange)
• Add extensions for distributions (#136) (@findmyway)

v0.4.4

ReinforcementLearningCore v0.4.4

Diff since v0.4.3

Merged pull requests:

• fixed typo. (#125) (@norci)

v0.4.3

ReinforcementLearningCore v0.4.3

Diff since v0.4.2

Closed issues:

• Add ElasticArray as a container in Trajectory (#112)

Merged pull requests:

• revert autocheck (#120) (@findmyway)
• add ElasticArray as container (#121) (@findmyway)
• CompatHelper: add new compat entry for "ElasticArrays" at version "1.2" (#122) (@github-actions[bot])

v0.4.2

ReinforcementLearningCore v0.4.2

Diff since v0.4.1

Closed issues:

• Implemente StatsBase.wsample so can sample from action space with certain probabilitly like in a policy (#94)
• Remove extract_experience (#108)

Merged pull requests:

• support mask for more explorers (#109) (@findmyway)
• Add reset method for StackFrames (#113) (@findmyway)
• add TabularLearner (#115) (@findmyway)
• Multi agent related changes (#116) (@findmyway)
• Multi agent changes (#118) (@findmyway)

v0.4.1

ReinforcementLearningCore v0.4.1

Diff since v0.4.0

Merged pull requests:

• CompatHelper: bump compat for "FillArrays" to "0.9" (#105) (@github-actions[bot])
• Trajectory redesign (#106) (@findmyway)

v0.4.0

ReinforcementLearningCore v0.4.0

Diff since v0.3.3

Closed issues:

• Adding a render hook or a render option in run() (#73)
• Reward transformations (#85)
• Ambiguity in update! for CircularArrayBuffer (#91)

Merged pull requests:

• Update dependencies (#90) (@findmyway)
• Update compat (#102) (@findmyway)

v0.3.3

ReinforcementLearningCore v0.3.3

Diff since v0.3.2

Merged pull requests:

• Extend existing hook to recognise RewardOverriddenObs (#86) (@findmyway)

v0.3.2

ReinforcementLearningCore v0.3.2

Diff since v0.3.1

Closed issues:

• Type definition in update function of CircularArrayBuffer structure (#76)

Merged pull requests:

• add orthogonal init function (#77) (@findmyway)
• CompatHelper: bump compat for "Adapt" to "2.0" (#79) (@github-actions[bot])
• update doc (#80) (@findmyway)
• CompatHelper: bump compat for "GPUArrays" to "4.0" (#81) (@github-actions[bot])
• Fix some corner cases when sampling a SumTree (#83) (@findmyway)