jens and Claude Sonnet 5
455ea692f5
Make --train, --human, --load plain flags (no value required)
...
Uses argparse.BooleanOptionalAction so these behave like --verbose
(e.g. "--human" instead of "--human True"), while keeping their
current defaults (False, False, True) and adding --no-train/--no-human/
--no-load counterparts.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com >
Claude-Session: https://claude.ai/code/session_01PQS7ZHqxQfy5DhKcz2gHau
2026-07-15 21:16:39 +02:00
jens and Claude Sonnet 5
2f56608c54
Make --num control the number of play games, not just training
...
--num was declared as "Number of games to play" but the play() call
used a hardcoded 10. Also give --num an int type so it works when
passed on the CLI (range() would otherwise crash on the raw string),
and make --verbose a store_true flag.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com >
Claude-Session: https://claude.ai/code/session_01PQS7ZHqxQfy5DhKcz2gHau
2026-07-15 21:14:26 +02:00
jens and Claude Sonnet 5
2fbbd0c712
Add README documenting project structure and CLI usage
...
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com >
Claude-Session: https://claude.ai/code/session_01PQS7ZHqxQfy5DhKcz2gHau
2026-07-15 21:10:33 +02:00
jens
aec0ec623e
added verbos argument
2025-03-12 08:27:09 +01:00
jens
1941061ed4
take program flags from command line
2025-03-12 08:22:11 +01:00
jens
f70cf1fc1e
- added reward sum and reward mean
2024-06-13 08:43:08 +02:00
jens
c50553c94a
fixed endgame
2024-06-11 22:05:06 +02:00
jens
7edfa627d0
Changed training to update values-file during training
2024-06-11 21:49:49 +02:00
jens
7a1156125c
- bugfix not rewarding properly
...
- bugfix not set value properly after learn
2024-06-11 21:15:55 +02:00
jens
e8d07f9d84
- fixed human player
2024-06-11 20:18:37 +02:00
jens
cf7379bf50
add learning after episode
2024-06-11 19:55:51 +02:00
jens
53e4d91ff8
- refactored learning function
...
- remove state_last, use state instead
- added episode history
2024-06-11 09:00:41 +02:00
jens
73f59f25fc
Refactored
2024-06-10 21:08:41 +02:00
jens
3064a8f0b2
No sampling
2024-06-10 21:08:27 +02:00
jens
e47959373b
Refactored
2024-06-10 21:07:38 +02:00
jens
61d1ab0431
Refactored
2024-06-10 20:48:50 +02:00
jens
285549568d
Re-indent with tabs
2024-06-10 20:37:26 +02:00
jens
d5b86b057e
Added human player
2024-06-10 08:37:21 +02:00
jens
59d1d28357
added report counter
2024-06-09 19:00:21 +02:00
jens
c1f64943c3
- simplified debug flags
...
- added params class
- cleaned up
2024-06-09 18:43:42 +02:00
jens
64960873c8
explorative moves which are identical to best moves estimation will learn
2024-06-09 17:33:10 +02:00
jens
19c972bc7d
test sample() only when main
2024-06-09 17:30:24 +02:00
jens
7cad1442df
Save values to file
2024-06-09 17:21:35 +02:00
jens
336a9b2cc9
refactored files
2024-06-09 17:03:17 +02:00
jens
aaff8e44e0
Fixed awarding and penalty at the end of each game
2024-06-09 16:42:32 +02:00
jens
1cd1fb729e
- refacored state
2024-06-09 15:03:35 +02:00
jens
ad32b75e50
Added silent training
2024-06-09 00:37:04 +02:00
jens
a812aebb77
Added probability paramter for exploration
2024-06-09 00:29:43 +02:00
jens
53608e96f0
Fixed sample
2024-06-08 23:34:09 +02:00
jens
cdc71b3eab
Fixed sample
2024-06-08 23:31:37 +02:00
jens
b9f83c26a1
add training without debug
2024-06-08 21:09:06 +02:00
jens
8cdb816409
Refactored
2024-06-08 21:00:00 +02:00
jens
f97c3b6f56
- refactored
2024-06-08 20:47:30 +02:00
jens
5b7430f208
- control debug output
...
- fixed last state
- refactored
2024-06-08 20:24:38 +02:00
jens
3a0246d0c9
- added idea project
...
- improved formatting
- improved state print
2024-06-08 18:55:19 +02:00
jens
da12213f95
Initial commit
2024-06-08 17:54:04 +02:00
jensandClaude Sonnet 5