Draws the shortest path from start to goal assuming a standard discrete action space.

Return the url and etag (which may be ``None``) stored for `filename`. Raise ``EnvironmentError`` if `filename` or its stored metadata do not

transformer processing

Args: **p_mask**: (`optional`) ``torch.FloatTensor`` of shape `(batch_size, seq_len)` invalid position mask such as query

Args: One of ``start_states``, ``start_positions`` should be not None. If both are set, ``start_positions`` overrides ``start

Args: One of ``start_states``, ``start_positions`` should be not None. If both are set, ``start_positions`` overrides

hidden_states: float Tensor in shape [bsz, seq_len, hidden_size], the hidden-states of the last layer. token_ids: [optional] index of the

hidden_states: (N, L_{hidden}, D) attention_mask: (N, H, L_{hidden}, L_{hidden})

Tensor sizes after computation: instruction: [batch_size x seq_length] lengths: [batch_size] hidden_state

Args: observations: [BATCH, HEIGHT, WIDTH, CHANNEL] Returns: [BATCH, OUTPUT_SIZE]

r"""Sends RGB observation through the TorchVision ResNet50 pre-trained on ImageNet. Sends through fully connected layer, activates, and

r"""Sends RGB observation through the TorchVision ResNet50 pre-trained on ImageNet. Sends through fully connected layer, activates, and

r""" Instantiate a PretrainedConfig from a pre-trained model configuration. Params: **pretrained_model_name_or_path**: either: