CosmicClash Training Bot d1ccf4dbc7 chore(training): generation 5 progress after 20260810-0211-gen5-s4-handling-retry2

2026-08-10 08:22:52 +01:00

..

checkpoints

chore(training): Add 20260810-0211-gen5-s4-handling-retry2 checkpoints, logs, and exported policy

2026-08-10 08:21:17 +01:00

logs

chore(training): Add 20260810-0211-gen5-s4-handling-retry2 checkpoints, logs, and exported policy

2026-08-10 08:21:17 +01:00

cosmic_env.py

feat(training): curriculum generation 4 — MultiDiscrete action space redesign

2026-08-04 23:27:57 +01:00

curriculum_state_gen1.json

…

curriculum_state_gen2.json

feat(training): Replace all-or-nothing unmask with a gradual ramp

2026-07-31 21:47:53 +01:00

curriculum_state_gen3.json

chore(training): scrap generation 1-3 training data for generation 4

2026-08-04 23:28:57 +01:00

curriculum_state.json

chore(training): curriculum progress after 20260805-1926-curric-s2-selfplay

2026-08-06 19:39:02 +01:00

curriculum.py

feat(training): curriculum generation 4 — MultiDiscrete action space redesign

2026-08-04 23:27:57 +01:00

curriculum.sh

…

eval_history.json

chore(training): generation 5 progress after 20260810-0211-gen5-s4-handling-retry2

2026-08-10 08:22:52 +01:00

evaluate.py

feat(training): pair evaluations across physical sides

2026-08-08 14:55:12 +01:00

export_linux.sh

feat(*): Add exported Linux binary training path for faster parallel instances

2026-07-28 20:56:14 +01:00

export_policy.py

feat(training): curriculum generation 4 — MultiDiscrete action space redesign

2026-08-04 23:27:57 +01:00

generation5_state.json

chore(training): generation 5 progress after 20260810-0211-gen5-s4-handling-retry2

2026-08-10 08:22:52 +01:00

generation5.py

fix(training): correct non-forward penalty math and add a grounding incentive

2026-08-09 13:23:00 +01:00

generation5.sh

feat(training): add generation 5 curriculum

2026-08-08 14:56:17 +01:00

next_long_run.sh

…

next_run.sh

…

requirements.txt

feat(training): curriculum generation 4 — MultiDiscrete action space redesign

2026-08-04 23:27:57 +01:00

run_training.sh

feat(training): curriculum generation 4 — MultiDiscrete action space redesign

2026-08-04 23:27:57 +01:00

setup_linux.sh

feat(*): Add exported Linux binary training path for faster parallel instances

2026-07-28 20:56:14 +01:00

start_training.sh

…

test_action_space.py

fix(ai): map team-relative rotation actions

2026-08-08 14:54:27 +01:00

test_evaluate.py

feat(training): pair evaluations across physical sides

2026-08-08 14:55:12 +01:00

test_generation5.py

feat(training): add generation 5 curriculum

2026-08-08 14:56:17 +01:00

train.py

fix(training): correct non-forward penalty math and add a grounding incentive

2026-08-09 13:23:00 +01:00