Returns: bool: whether the arguments contain :class:`Configuration` and should be considered forwarded to from_config.

Adjust model's first convolution layer to in_channels, if in_channels not equals 3. Args: in_channels (int): Input data's channel

Validates model name. Args: model_name (str): Name for efficientnet. Returns: bool: Is a valid name or not.

assume that the distance between text and video in the same pair is smaller than they each compared with other modality in other pairs

Return an activation function given a string

Use `from_config` to obtain explicit arguments. Returns: dict: arguments to be used for :func:`cls.__init__`

Returns possible attribute that can point to image id Args: item (Object): Object from the DB Returns: List[

Args: video_path(str): video path. start(float): start timestamp of video for sampling frames duration(f

Calculate mce accuracy & recall.

Preprocess an image or batch of images. Copy of the `preprocess` method from `CLIPImageProcessor`. Args: images (`ImageI

Print sample-wise metrics and batch-wise losses/metrics, used in validating a whole val/test set.

Truncates a sequence pair in place to the maximum length.

Build configuration from command line.

AutoTokenizer needs transformer config to decide which tokenizer instance to initialize: * If we have downloaded the pretrained model, `token

Returns list of utf-8 byte and a corresponding list of unicode strings. The reversible bpe codes work on unicode strings. This means you

Abstract method to be implemented by the child class. Takes in a ``SampleList`` and a dict returned by model as output and returns bac

Performs a HEAD request to check if the URL / Google Drive ID is live.

Checksum on a given file. Args: download_path (string): path to the downloaded file.

Override this in case you want to clean the config you updated earlier in update_registry_for_model

Args: images: a tuple or list of `torch.Tensor`, each of shape (Hi, Wi) or (C_1, ..., C_K, Hi, Wi) where K >= 1. The Tensors

Find the most likely tag sequence using Viterbi algorithm. Args: emissions (`~torch.Tensor`): Emission score tensor of size