rl_salamandra_alignment.trl_scripts package

Submodules

rl_salamandra_alignment.trl_scripts.alignprop module

rl_salamandra_alignment.trl_scripts.bco module

rl_salamandra_alignment.trl_scripts.chat module

rl_salamandra_alignment.trl_scripts.cpo module

rl_salamandra_alignment.trl_scripts.ddpo module

rl_salamandra_alignment.trl_scripts.dpo module

rl_salamandra_alignment.trl_scripts.dpo_online module

rl_salamandra_alignment.trl_scripts.dpo_vlm module

rl_salamandra_alignment.trl_scripts.gkd module

rl_salamandra_alignment.trl_scripts.grpo module

rl_salamandra_alignment.trl_scripts.kto module

rl_salamandra_alignment.trl_scripts.nash_md module

rl_salamandra_alignment.trl_scripts.orpo module

rl_salamandra_alignment.trl_scripts.reward_modeling module

rl_salamandra_alignment.trl_scripts.sft module

rl_salamandra_alignment.trl_scripts.sft_video_llm module

rl_salamandra_alignment.trl_scripts.sft_vlm module

rl_salamandra_alignment.trl_scripts.xpo module

Module contents

TRL scripts for Reinforcement Learning Algorithms

rl_salamandra_alignment.trl_scripts.get_script_path(algorithm_name)[source]

Get the absolute path to a python script file for a Reinforcement Learning algorithmn

Parameters:: algorithm_name (TRL_Algorithms) – Name Reinforcement Learning algorithm
Returns:: Path to python script
Return type:: str