 fanasina
|
93a89eaa99
|
[update] use tensor contract prod Opt0 in learn to drive algo
|
2026-02-11 00:30:13 +01:00 |
|
|
fanasina
|
d8ae2729df
|
[update] learn to drive: decrease learning rate when very good rewards
|
2026-02-11 00:28:28 +01:00 |
|
|
fanasina
|
5bdb030587
|
add char sep to nameid and filename, org in y_file_handler.c
|
2025-12-21 20:07:50 +01:00 |
|
|
fanasina
|
2950b20a8f
|
learn_to_drive: action repeated too much, increase exploration factor
|
2025-12-15 21:13:52 +01:00 |
|
|
fanasina
|
ddb9e44536
|
learn_to_drive: debug free qparams call when free rlagent
|
2025-12-15 12:35:07 +01:00 |
|
|
fanasina
|
77ca0ae5c7
|
learn_to_drive: add debug print when free qlparam
|
2025-12-15 12:29:46 +01:00 |
|
|
fanasina
|
0e80467c19
|
qlearn: add func caller name to store result more verbose, and add record result if ending before done
|
2025-12-15 11:15:51 +01:00 |
|
|
fanasina
|
3d54639d4c
|
debug : nan, it's the learning rate too high,s try to find optimum: 0.001
|
2025-12-13 05:00:23 +01:00 |
|
|
fanasina
|
32207170f6
|
learn_to_drive: add up threshold in reLU to avoid NaN in neuron weight calc
|
2025-11-28 20:53:04 +01:00 |
|
|
fanasina
|
3daabbe634
|
y_nnn learn_to_drive: try to minimise no change action! search optimal params
|
2025-11-25 15:06:48 +01:00 |
|
|
fanasina
|
b505673bda
|
y_nnn: launch learn when rlAgent ready, add thread attribut in networks to follow thread to avoid mem leak
|
2025-11-23 18:42:45 +01:00 |
|
|
fanasina
|
c60e41cf0f
|
y_socket: transfert kill to extern, y_nnn: debug leak mem
|
2025-11-23 13:30:01 +01:00 |
|
|
fanasina
|
ba0a95a5b1
|
learn_to_drive: change learn_to_drive signature to be launch in a thread
|
2025-11-19 22:48:25 +01:00 |
|
|
fanasina
|
ea26f5f3bc
|
y_nnn: WORK IN PROGRESS... have to save
|
2025-11-16 00:01:16 +01:00 |
|
|
fanasina
|
a59b9a5538
|
y_nnn : change learn_to_drive and add some bash print
|
2025-11-07 12:55:43 +01:00 |
|
|
fanasina
|
151daf548d
|
y_socket: add help and kill cmd, update usage_cmdl
|
2025-10-23 11:02:35 +02:00 |
|
|
fanasina
|
244ac9cd1d
|
deepQlearning: add libdeepqlearn.so, debug some issue with -Werror -fpic
|
2025-10-21 08:34:25 +02:00 |
|
|
fanasina
|
52e6353469
|
learnig: add ending and fix leak memory
|
2025-05-12 15:16:19 +02:00 |
|
|
fanasina
|
bdaf894aea
|
add sockets
|
2024-12-24 23:15:44 +01:00 |
|
|
fanasina
|
e64cef1688
|
add some test deep reinforcement learning
|
2024-07-18 02:10:41 +02:00 |
|
|
fanasina
|
aac7434346
|
Trying to fix nan output of RL by using relu with upperbound
|
2024-07-16 12:13:05 +02:00 |
|
|
fanasina
|
0c9813beca
|
Add some MACRO in neuron_t and debug deepQlearning
|
2024-07-09 17:38:58 +02:00 |
|
|
fanasina
|
824396f901
|
second try, not yet working
|
2024-06-21 11:37:43 +02:00 |
|
|
fanasina
|
582d7a6a70
|
first try qdeep learning
|
2024-06-21 00:28:50 +02:00 |
|
|
fanasina
|
e96304bee7
|
wip deepQlearning
|
2024-06-14 23:07:39 +02:00 |
|
|
fanasina
|
474af56377
|
using tensor in vehicle
|
2024-06-14 15:59:47 +02:00 |
|
|
fanasina
|
9927d6642c
|
modify COMPARE_N in tool, modify attribute of vehicle by using tensor
|
2024-06-13 23:35:25 +02:00 |
|
|
fanasina
|
13f91583bb
|
update learn to drive and Makefile
|
2024-06-12 00:24:17 +02:00 |
|
|
fanasina
|
fca991eb37
|
create deepReinforcementLearning repo, test print vehicle
|
2024-06-11 22:24:59 +02:00 |
|