harl.common package

Subpackages

• harl.common.buffers package
  • Submodules
  • harl.common.buffers.off_policy_buffer_base module
    • OffPolicyBufferBase
      • OffPolicyBufferBase.get_mean_rewards()
      • OffPolicyBufferBase.insert()
      • OffPolicyBufferBase.next()
      • OffPolicyBufferBase.sample()
      • OffPolicyBufferBase.update_end_flag()
  • harl.common.buffers.off_policy_buffer_ep module
    • OffPolicyBufferEP
      • OffPolicyBufferEP.next()
      • OffPolicyBufferEP.sample()
      • OffPolicyBufferEP.update_end_flag()
  • harl.common.buffers.off_policy_buffer_fp module
    • OffPolicyBufferFP
      • OffPolicyBufferFP.next()
      • OffPolicyBufferFP.sample()
      • OffPolicyBufferFP.update_end_flag()
  • harl.common.buffers.on_policy_actor_buffer module
    • OnPolicyActorBuffer
      • OnPolicyActorBuffer.after_update()
      • OnPolicyActorBuffer.feed_forward_generator_actor()
      • OnPolicyActorBuffer.insert()
      • OnPolicyActorBuffer.naive_recurrent_generator_actor()
      • OnPolicyActorBuffer.recurrent_generator_actor()
      • OnPolicyActorBuffer.update_factor()
  • harl.common.buffers.on_policy_critic_buffer_ep module
    • OnPolicyCriticBufferEP
      • OnPolicyCriticBufferEP.after_update()
      • OnPolicyCriticBufferEP.compute_returns()
      • OnPolicyCriticBufferEP.feed_forward_generator_critic()
      • OnPolicyCriticBufferEP.get_mean_rewards()
      • OnPolicyCriticBufferEP.insert()
      • OnPolicyCriticBufferEP.naive_recurrent_generator_critic()
      • OnPolicyCriticBufferEP.recurrent_generator_critic()
  • harl.common.buffers.on_policy_critic_buffer_fp module
    • OnPolicyCriticBufferFP
      • OnPolicyCriticBufferFP.after_update()
      • OnPolicyCriticBufferFP.compute_returns()
      • OnPolicyCriticBufferFP.feed_forward_generator_critic()
      • OnPolicyCriticBufferFP.get_mean_rewards()
      • OnPolicyCriticBufferFP.insert()
      • OnPolicyCriticBufferFP.naive_recurrent_generator_critic()
      • OnPolicyCriticBufferFP.recurrent_generator_critic()
  • Module contents

Submodules

harl.common.base_logger module

Base logger.

class harl.common.base_logger.BaseLogger(args, algo_args, env_args, num_agents, writter, run_dir)

Bases: object

Base logger class. Used for logging information in the on-policy training pipeline.

close()

Close the logger.

episode_init(episode)

Initialize the logger for each episode.

episode_log(actor_train_infos, critic_train_info, actor_buffer, critic_buffer)

Log information for each episode.

eval_init()

Initialize the logger for evaluation.

eval_init_off_policy(total_num_steps)

eval_log(eval_episode)

Log evaluation information.

eval_per_step(eval_data)

Log evaluation information per step.

eval_thread_done(tid)

Log evaluation information.

get_task_name()

Get the task name.

init(episodes)

Initialize the logger.

log_env(env_infos)

Log environment information.

log_train(actor_train_infos, critic_train_info)

Log training information.

per_step(data)

Process data per step.

harl.common.valuenorm module

ValueNorm.

class harl.common.valuenorm.ValueNorm(input_shape, norm_axes=1, beta=0.99999, per_element_update=False, epsilon=1e-05, device=device(type='cpu'))

Bases: Module

Normalize a vector of observations - across the first norm_axes dimensions

denormalize(input_vector)

Transform normalized data back into original distribution

normalize(input_vector)

running_mean_var()

Get running mean and variance.

update(input_vector)

Module contents