Leonora Tindall
49b720baec
Use a normal distribution for mutations
...
Rather than StandardNormal, weights are generated with a Normal
distribution centered on 0 with a standard deviation of 0.75,
and mutations are performed by adding a number from that distribution to
the existing weight, rather than replacing the weight entirely.
2023-04-03 22:32:53 -05:00
b7ad50da37
activ func
2023-01-12 15:25:26 +05:30
8d317346f8
hlayers, mut_rate ui bridged
2023-01-12 03:01:03 +05:30
39941c8425
side ui
2023-01-12 00:26:03 +05:30
b9bc8ae08d
debug, selection
2023-01-10 00:55:18 +05:30
7110d063b4
speed control
2023-01-09 02:10:34 +05:30
31be57ab37
load/save models
2023-01-09 01:39:10 +05:30
7255bd617d
nn ui text
2023-01-09 00:56:00 +05:30
7e3acb4d57
player visual
2023-01-09 00:37:35 +05:30
7dea9e0406
ui bars
2023-01-08 03:29:59 +05:30
3ce1fdb957
visuals
2023-01-07 04:17:57 +05:30
02df5694f9
import/export models
2023-01-06 14:40:29 +05:30
b9cfedc00d
cleanup
2023-01-06 02:03:25 +05:30
8196383515
finally works T_T im ded
2023-01-05 01:27:59 +05:30
1c29b6d419
bug fixes
2022-10-23 01:10:19 +05:30
db2df770c6
fixed bugs
2022-10-14 16:27:58 +05:30
1f0b6e06e0
genetic algorithm
2022-10-11 00:06:14 +05:30
4084a43acc
population
2022-10-10 01:41:24 +05:30
91bc4e27a4
crossover, mutation
2022-10-10 01:16:27 +05:30
44e2cb36f4
feed_forward: Vec -> Vec
2022-10-10 00:03:00 +05:30
80fb255c36
neural network, feed forward
2022-10-09 23:44:22 +05:30
Leonora Tindall
sparshg