timandClaude Fable 5 9a750756c5 RL: pure-Python inferencia natrenovanej siete
py -m rl.export vyexportuje checkpoint do rl/weights/neural-bot.json
(bit-exact float32, 1.3 MB) a rl/pure_net.py ho hra bez torch/numpy
(stdlib forward pass, ~16 ms/tah). Testy parity: logity aj akcie sa
zhoduju s torch, identicke trajektorie celych kol. Natrenovany model:
6.7-6.9 b/kolo proti vsetkym baseline-om (heuristika prekonana).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-07 18:49:55 +02:00
2022-07-05 15:58:23 +02:00
2026-07-04 13:45:41 +02:00
2026-07-07 18:49:55 +02:00
2026-07-01 20:05:40 +02:00
S
Description
kartova hra bridzik
1.6 MiB
Languages
Python 100%