Commit Graph

28 Commits

Author SHA1 Message Date
e085238e2f V2.1.1
Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>

Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>
2023-08-30 21:19:36 +02:00
9c2caaa444 V2.1.1
Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>

Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>
2023-08-30 13:47:26 +02:00
e32b580166 Update train.py 2023-08-29 15:09:59 +02:00
885495fcd1 Add files via upload 2023-08-29 14:19:18 +02:00
8329e282d3 Merge remote-tracking branch 'Snuf/master' into pwnagotchi-torch
# Conflicts:
#	bin/pwnagotchi
#	builder/data/usr/bin/pwnlib
#	pwnagotchi/ai/__init__.py
2023-07-24 21:20:37 +02:00
27da741005 Edited to make AI work
Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>
2023-07-24 19:47:15 +02:00
702d32292f Update __init__.py 2023-07-23 12:28:20 +02:00
ce0d99c46d Minor edits
Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>
2023-07-21 10:25:09 +02:00
8a00c1835b Removed unused element
Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>
2023-07-21 00:03:38 +02:00
5da9f3598a Changed for torch installation
Signed-off-by: Jeroen Oudshoorn <oudshoorn.jeroen@gmail.com>
2023-07-20 19:55:56 +02:00
cd92491e40 move brain.nn backup into ai_worker before training session instead of backing up on every save. keep the "previous trained" brain 2023-07-04 13:18:18 -07:00
c52220d98e added support for stable_baselines3 AI backend. stable_baselines3 uses
pytorch instead of tensorflow
2023-06-18 18:56:06 -07:00
d2227b939d changed cutoff for "extended channels" from 140 to 150 to alleviate "shape into shape" ai error
- more of a bandaid than a fix
2023-06-18 18:28:57 -07:00
d2c52e2673 save backups of brain files when saving 2023-05-19 11:01:08 -07:00
4b89a3c8dd use a tag on cpu_load for more accurate load over the last epoch 2023-05-19 10:59:41 -07:00
c90855840c keep old version as backup when saving files 2023-04-27 15:00:58 -07:00
1f2dd73976 Big update 2020-04-13 17:16:24 +02:00
a8c07ba997 fix: supporting channels greater than 140 for 5g (closes #583) 2019-11-13 01:32:05 +01:00
2c9f54567f misc: small fix or general refactoring i did not bother commenting 2019-10-25 13:21:08 +02:00
8b00e0ae10 misc: small fix or general refactoring i did not bother commenting 2019-10-25 12:17:44 +02:00
7954bb5fcf misc: small fix or general refactoring i did not bother commenting 2019-10-25 12:15:56 +02:00
06d8cc63fb misc: added debug logs for AI loading times 2019-10-25 11:40:58 +02:00
ec430a5cba misc: small fix or general refactoring i did not bother commenting 2019-10-24 19:54:55 +02:00
dc5a626bd5 misc: small fix or general refactoring i did not bother commenting 2019-10-24 15:34:52 +02:00
608aad4820 new: observing and reporting total number of peers met per each epoch 2019-10-24 15:18:04 +02:00
36c3ea5bbc new: new grateful status that can override sad/bored/lonely if units with a strong bond are nearby 2019-10-23 18:19:43 +02:00
61da16ed91 fix: when the AI is ready the mode label reports AI 2019-10-07 01:31:37 +02:00
cbf7511e17 misc: refactored code to work as a setup.py based package
Signed-off-by: Simone Margaritelli <evilsocket@gmail.com>
2019-10-05 23:23:31 +02:00