Create a semi-unique experiment name that has a timestamp :param exp_prefix: :param exp_id: :return:

Creates and returns a unique log directory. :param exp_prefix: All experiments with this prefix will have log directories be under this

Convert each value in the dictionary into a JSON'able primitive. :param d: :return:

The following value for the following keys will be a 2D array, with the first dimension corresponding to the time dimension. - observati

Return a value for the hyperparameter

:param latent_distribution_params: :return: get latents from latent distribution params

The intended use is to have a subprocess serialize/copy a SharedObsDictRelabelingBuffer instance and call init_from on the in

Load a single replay buffer.

Assign flags for a quick run of FixedReplay agent.

:param latent_distribution_params: :return: latents

Call this between calls to separate experiments. :return:

Assume goal contains both image_desired_goal and any goals required for wrapped envs :param goal: :return:

Set the seed for all the possible random number generators. :param seed: :return: None

Image env and pickup env will have presampled goals. VAE wrapper should encode whatever presampled goal is sampled.

Stack multiples obs/actions/etc. from different paths :param paths: List of paths, where one path is something returned from the rollout

Split "s3://foo/bar/baz" into "foo" and "bar/baz"

Let the replay buffer know that the episode has terminated in case some special book-keeping has to happen. :return:

Gets the priorities for the given indices. Args: indices: tf.Tensor with dtype int32 and shape [n]. Returns: priorities: tf.Tens

Sets the priorities for the given indices. Args: indices: tf.Tensor with dtype int32 and shape [n]. priorities: tf.Tensor with dtype

Verify that files have been created.