r"""Main method of trainer evaluation. Calls _eval_checkpoint() that is specified in Trainer class that inherits from BaseRLTrainer or

r"""This control method is called from ``Env`` on each ``step``.

Extracts instruction tokens from an instruction sensor if the tokens exist and are in a dict structure.

r"""Create VectorEnv object with specified config and env class type. To allow better performance, dataset are split into small ones for eac

B x [T, ...] torch tensors

Conv1D layer as defined by Radford et al. for OpenAI GPT (and also used in GPT-2) Basically works like a Linear layer but the weights are

Prune a linear layer (a model parameters) to keep only entries in index. Return the pruned layer as a new layer with requires_grad=True.

wrap action, get obs if video_option

Return an activation function given a string

r"""Instantiate a pretrained pytorch model from a pre-trained model configuration. The model is set in evaluation mode by default using `

Return gridworld index of realworld coordinates assuming top-left corner is the origin. The real world coordinates of lower left corner are (c

angle: 0 ~ 360 degree

Args: equ_h: (int) the height of the generated equirect equ_w: (int) the width of the generated equirect

r"""Helper method to get the scene name from an episode. Assumes the scene_id is formated /path/to/<scene_name>.<ext>

mask from (N, L) into (N, 1(H), 1(L), L) and make it negative

r"""Create a unified config with default values. Initialized from the habitat_baselines default config. Overwritten by values from `config_pat

r"""Heuristic to determine if a slurm job is a batch job or not. Batch jobs will have a job name that is not a shell unless the user specifically

Make sure we are sharing the input and output embeddings. Export to TorchScript can't handle parameter sharing so we are cloning them ins

r"""Evaluates a single checkpoint. Args: checkpoint_path: path of checkpoint writer: tensorboard writer object

Tie or clone module weights depending of weither we are using TorchScript or not

r"""Appends text underneath an image of size (height, width, channels). The returned image has white text on a black background. Uses textwrap to

Given something that might be a URL (or might be a local path), determine which. If it's a URL, download the file and cache it, and retur

r"""Given the output of the TopDownMap measure, colorizes the map, draws the agent, and fits to a desired output height :param info: The outp

https://stackoverflow.com/questions/26690932/opencv-rectangle-with-dotted-or-dashed-lines style: "dotted", "dashed", or "filled"

Generate video according to specified information. Using a custom verion instead of Habitat's that passes FPS to video maker. Args: v

r"""Return a sorted list of scenes

Initialize the weights.

Creates a rudimentary frame for a panoramic observation. Includes RGB, depth, and a top-down map. TODO: create a visually-pleasing stitched pa

Transforms model's predictions to semantically meaningful values. Args: raw: [num_rays, num_samples along ray, 4]. Prediction from model

Split a full s3 path into the bucket name and path.