CosmicClash Training Bot f08ecbe651 chore(training): Add 20260813-0134-gen5-s4-handling-retry2 checkpoints, logs, and exported policy

2026-08-13 07:44:31 +01:00

..

checkpoints

chore(training): Add 20260813-0134-gen5-s4-handling-retry2 checkpoints, logs, and exported policy

2026-08-13 07:44:31 +01:00

logs

chore(training): Add 20260813-0134-gen5-s4-handling-retry2 checkpoints, logs, and exported policy

2026-08-13 07:44:31 +01:00

cosmic_env.py

…

curriculum_state_gen1.json

…

curriculum_state_gen2.json

…

curriculum_state_gen3.json

…

curriculum_state.json

…

curriculum.py

…

curriculum.sh

…

eval_history.json

chore(training): generation 5 progress after 20260812-1922-gen5-s4-handling-retry1

2026-08-13 01:34:15 +01:00

evaluate.py

…

export_linux.sh

…

export_policy.py

…

generation5_state.json

chore(training): generation 5 progress after 20260812-1922-gen5-s4-handling-retry1

2026-08-13 01:34:15 +01:00

generation5.py

fix(training): make uprightness a multiplier on the forward-approach reward

2026-08-12 13:04:20 +01:00

generation5.sh

…

next_long_run.sh

…

next_run.sh

…

requirements.txt

…

run_training.sh

…

setup_linux.sh

…

start_training.sh

…

test_action_space.py

…

test_evaluate.py

…

test_generation5.py

…

train.py

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

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