feature(whl/pu): add jericho environments and related configs#307

Merged

puyuan1996 merged 28 commits intomainfrom dev-nlp-pu

Mar 18, 2025

Commits on Dec 1, 2024

init commit
‘whl’
committed
add bert encoding
‘whl’
committed
debug
‘whl’
committed
debug and polish env
‘whl’
committed
polish config
‘whl’
committed
update
‘whl’
committed

Commits on Dec 3, 2024

polish(pu): polish jericho config and pipeline
dyyoungg
committed

Commits on Dec 13, 2024

Commits on Dec 19, 2024

Commits on Dec 20, 2024

fix(pu): fix train_entry_time to the same in ddp train
puyuan1996
committed

Commits on Jan 7, 2025

polish(pu): polish jericho uz config

puyuan1996
authored and
puyuan
committed

Commits on Jan 8, 2025

feature(pu): add remove_stuck_actions option
puyuan
committed

Commits on Jan 14, 2025

polish(pu): set projection layer train after language embed model
PaParaZz1
committed

Commits on Jan 21, 2025

polish(pu): polish jericho_env and add jericho_ppo_config
PaParaZz1
committed

Commits on Jan 22, 2025

sync code
PaParaZz1
committed

Commits on Feb 8, 2025

fix(pu): fix gradient accumulation_steps option
puyuan
committed

Commits on Feb 12, 2025

fix(pu): fix action_mask all-zero bug
puyuan
committed

Commits on Feb 13, 2025

fix(pu): fix lr target_model_update bug when accumulation_steps>1
puyuan
committed

Commits on Mar 16, 2025