v0.9.4

What's Changed

• Fix network propagation bug and improve test coverage by @nleck in #184

Full Changelog: v0.9.3...v0.9.4

v0.9.3

What's Changed

• Added logo to Readme by @nleck in #178
• 179 if bias back propagation by @nleck in #180
• Fix type casting issue in Network.ts and Node.ts by @nleck in #181
• Fix trace dir by @nleck in #182

Full Changelog: v0.9.2...v0.9.3

Second pre-release

What's Changed

• 176 if back propagation by @nleck in #177

Full Changelog: v0.9.1...v0.9.2

Pre-release initial version

Pre-release of initial version

What's Changed

• Create LICENSE by @nleck in #1
• Deno worker by @nleck in #2
• refactor to typescript by @nleck in #3
• If else by @nleck in #4
• improve performance by @nleck in #5
• improved performance by @nleck in #6
• fixes. by @nleck in #7
• fix IF by @nleck in #8
• fixed others. by @nleck in #9
• docs by @nleck in #10
• Allow self by @nleck in #11
• IF fixed. by @nleck in #12
• take the connections as well as the node by @nleck in #13
• improving from by @nleck in #14
• no need to export input nodes by @nleck in #15
• Crisper by @nleck in #16
• fix bias by @nleck in #18
• fixed NaN by @nleck in #19
• fixed by @nleck in #20
• Islands by @nleck in #21
• added new cost by @nleck in #22
• fixing infinite by @nleck in #23
• Fix infinite by @nleck in #24
• Cross over mandates tags. by @nleck in #25
• Keep crispr by @nleck in #26
• increase errro by @nleck in #27
• added twelve steps by @nleck in #28
• -1->1 by @nleck in #29
• fixed cache issue by @nleck in #30
• fix tags by @nleck in #31
• Fixed leaking of state. by @nleck in #32
• improve performance by @nleck in #33
• Relative cross by @nleck in #34
• Implemented compact of network. by @nleck in #35
• Remove for each by @nleck in #36
• added diagnostics by @nleck in #37
• Slight improvement in performance. by @nleck in #38
• Compact nodes that only point to themselves by @nleck in #39
• Trace aggregate3 by @nleck in #40
• fixed possible infinite loop by @nleck in #41
• Trace activation4 by @nleck in #42
• added logging of learnings by @nleck in #43
• Apply learnings by @nleck in #44
• add SUM by @nleck in #45
• fix infinite by @nleck in #46
• Added dispose by @nleck in #47
• fix upgrade by @nleck in #48
• Remove sum by @nleck in #49
• #52 export format by @nleck in #53
• remove UUID from network. by @nleck in #55
• #56v2 by @nleck in #57
• #56v2 by @nleck in #59
• Use array for activation by @nleck in #61
• 60 use an array for activation by @nleck in #62
• 56 write out statistics by @nleck in #63
• 60 v2 by @nleck in #65
• limit the change by @nleck in #67
• Pass the input size. by @nleck in #73
• Fixed upgrade. by @nleck in #76
• removed throw by @nleck in #79
• Check if hidden/output by @nleck in #80
• Always be training. by @nleck in #82
• De couple the training from scoring. by @nleck in #84
• Fixing errors. by @nleck in #86
• Fixed messages and terminate workers. by @nleck in #88
• Fixed multiple outputs. by @nleck in #92
• fixed breading with many outputs. by @nleck in #94
• added test by @nleck in #97
• tidy up by @nleck in #99
• implemented by @nleck in #101
• Handle dups in CISPR by @nleck in #103
• Add small random connections before training. by @nleck in #105
• 106 fine tune based on UUID not position by @nleck in #107
• tag trained by @nleck in #109
• Upgraded standard library. by @nleck in #111
• two steps forward one back. by @nleck in #112
• trying to fix totalDeltaWeight by @nleck in #114
• Always fine tune from a random creature. by @nleck in #116
• Fixed usage of undefined. by @nleck in #119
• 118 output UUID should be output 0 output 1 etc by @nleck in #120
• simplified. by @nleck in #122
• removed. by @nleck in #124
• Upgrade Deno version by @nleck in #126
• Added more activation functions. by @nleck in #128
• Performance improvements. by @nleck in #130
• performance improvements by @nleck in #132
• Trying to improve performance by @nleck in #134
• Added bench mark by @nleck in #136
• 68 simple2 by @nleck in #139
• improvements by @nleck in #141
• Increased to 24 by @nleck in #142
• trying to improve performance. by @nleck in #143
• reverted changes. by @nleck in #144
• fixed offset by @nleck in #145
• clean up by @nleck in #147
• Added unSquash and range by @nleck in #149
• added activation functons. by @nleck in #151
• Upgrade std library by @nleck in #153
• 154 coverage by @nleck in #155
• Penalize exploding by @nleck in #157
• 159 test cases for expoding gradiants by @nleck in #160
• fixed quantum. by @nleck in #162
• upgrade to 0.208 by @nleck in #165
• 166 sort nodes correctly by @nleck in #167
• Upgrade by @nleck in #168
• Slight performance improvement by @nleck in #169
• Performance 20231224v2 by @nleck in #170
• Refactor training options and error handling by @nleck in #171
• fixed coverage by @nleck in #172
• use deno by @nleck in #173
• Rework back propagation by @nleck in #137
• Fix linting issues and update test configurations by @nleck in #175

Full Changelog: https://github.com/stSoftwareAU/NEAT-AI/commits/v0.9.1