All subclasses should override this method, if they need any operations before train. Args: runner (Runner): The runner o

Build a temporary directory containing many files to test ``FileClient.list_dir_or_file``. . \n | -- dir1 \n | -- | -- text3.txt \n

Build validation loop. Examples of ``loop``: # `ValLoop` will be used loop = dict() # custom validation

Returns the statistics for all submodules, separated out by operator type for each submodule. The operator handle determines the name

Convert various input in RGB order to normalized RGB matplotlib color tuples, Args: colors (Union[str, tuple, List[Union[str, tuple]]]

Copy a file src to dst and return the destination file. src and dst should have the same prefix. If dst specifies a directory, the fi

Get installed path of package. Args: package (str): Name of package. Example: >>> get_installed_path('mmcls') >>> '.

Interface to get the complexity of a model. The parameter `inputs` are fed to the forward method of model. If `inputs` is not specified, the

Check if a method of base class is overridden in derived class. Args: method (str): the method name to check. base_class (type):

Check whether a file path is a directory. Args: filepath (str or Path): Path to be checked whether it is a direct

Convert the label-format input to one-hot. Args: label (torch.Tensor): The label-format input. The format of item

Merge list into cfg_dict. Merge the dict parsed by MultipleKVAction into this cfg. Args: options (dict): dict of configs

Parses the raw outputs (losses) of the network. Args: losses (dict): Raw output of the network, which usually contain

int: A number to set random modules.

Special method for set union field, used as property.setter functions.

Access statistics method by name. Args: method_name (str): Name of method. Returns: Any: Depends on correspo

Update the log history. If the length of the buffer exceeds ``self._max_length``, the oldest element will be removed from the buffer.

Update parameters in :attr:`optimizer`. Args: loss (torch.Tensor): A tensor for back propagation. step_kwargs (dict):

Initialize BaseModule, inherited from `torch.nn.Module`

All subclasses should override this method, if they need any operations before each epoch. Args: runner (Runner): The run

Collect log information to compose a dict according to mode. Args: custom_cfg (List[dict]): A copy of ``self.custom_cfg`` with in

Get current epoch according to mode. Args: runner (Runner): The runner of the training/testing/validation process

Judge whether the position is in image. Args: position (np.ndarray): The position to judge which last dim must be

Parse window_size defined in custom_cfg to int value. Args: runner (Runner): The runner of the training/testing/validation

Depth-first search for the corresponding registry in its children. Note that the method only search for the corresponding registry from

Recursively set device for `BaseDataPreprocessor` instance. Args: device (torch.device): the desired device of the parameters and

Substitute environment variables in config with actual values. Sometimes, we want to change some items in the config with environment

Returns: list: Contains all keys in metainfo and data.

Build test log_processor. Examples of ``log_processor``: # `LogProcessor` will be used log_processor = dict()

Invoke ``evaluate`` method of each metric and collect the metrics dictionary. Args: size (int): Length of the entire vali

Build a Config instance from config text. Args: cfg_str (str): Config text. file_format (str): Config file format cor

Loop to ``full_init`` each dataset.

This method returns a handle that counts the number of activation from the output shape for the specified operation. Args: op_name (s

Get config from external package. Args: cfg_path (str): External relative config path. pretrained (bool): Whether to save pretrai

Get the git hash of the current repo. Args: fallback (str, optional): The fallback string when git hash is unavailable. Defau

Get the drawn image. The format is RGB. Returns: np.ndarray: the drawn image which channel is RGB.

Check package whether installed. Args: package (str): Name of package to be checked.

Returns True if Cambricon PyTorch and mlu devices exist.

Whether the input is an string instance. Note: This method is deprecated since python 2 is no longer supported.

Check whether a file path is a file. Args: filepath (str or Path): Path to be checked whether it is a file. Returns:

Load checkpoint from a file or URI. Args: model (Module): Module to load checkpoint. filename (str): Accept local filepath, URL,

Load checkpoint from given ``filename``. Args: filename (str): Accept local filepath, URL, ``torchvision://xxx``,

Returns: iterator: An iterator object whose element is (key, value) tuple pairs for ``metainfo``.

Returns: list: Contains all keys in metainfo_fields.

Enables the context for mixed precision training, and enables the context for disabling gradient synchronization during gradient accum

Prepare model and some components. Args: model (:obj:`torch.nn.Module` or dict): The model to be run. It can be a

Convert list of data sampled from dataset into a batch of data, of which type consistent with the type of each data_itement in ``data_batch``.

Write data to a given ``filepath`` with 'wb' mode. Note: ``put`` should create a directory if the directory of ``filepath``

Write bytes to a given ``filepath``. Args: obj (bytes): Data to be saved. filepath (str or Path): Path to write data.

Automatically scaling learning rate in training according to the ratio of ``base_batch_size`` in ``autoscalelr_cfg`` and real batch si

Scan a directory to find the interested files. Args: dir_path (str | :obj:`Path`): Path of the directory. suffix (str | tuple(str

Sets the epoch for this sampler. When :attr:`shuffle=True`, this ensures all replicas use a different random ordering for each epoch.

Set multi-processing related environment. Args: mp_start_method (str): Set the method which should be used to start child pro

Show the drawn image. Args: drawn_img (np.ndarray, optional): The image to show. If drawn_img is None, it will sh

Finds a loader that supports the given path. Falls back to the local loader if no other loader is found. Args: path (str)

Get current loop according to mode. Args: runner (Runner): The runner of the training/validation/testing process.

Convert global idx to local index. Args: idx (int): Global index of ``RepeatDataset``. Returns: Tuple[int, i

Initialize visualizers. Args: cfg (ConfigType): Config containing the visualizer information. Returns: Visua

Check if a module is a built-in module. Arg: module_name: name of module.

Record the scalar data to disk. Args: name (str): The scalar identifier. value (int, float, torch.Tensor, np.ndarray)

Record the scalars' data. Args: scalar_dict (dict): Key-value pair storing the tag and corresponding values.

Resume ema parameters from checkpoint. Args: runner (Runner): The runner of the testing process.

Show or Write the predicted results. Args: runner (Runner): The runner of the training process. batch_idx (int): The

All subclasses should override this method, if they need any operations after each training iteration. Args: runner (Runn

All subclasses should override this method, if they need any operations after each validation epoch. Args: runner (Runner

Sets how to determine the ancestor modules of an operator. Must be one of "owner" or "caller". * "caller": an operator belongs to all

Initialize the profiler. Through the runner parameter, the validity of the parameter is further determined.

If class attribute ``lazy`` is False, the LazyObject will be built and returned. Args: value (Any): The value to be built

Build a temporary directory containing many files to test ``FileClient.list_dir_or_file``. . \n | -- dir1 \n | -- | -- text3.txt \n

Build a temporary directory containing many files to test ``FileClient.list_dir_or_file``. . \n | -- dir1 \n | -- | -- text3.txt \n

Build a global asscessable Visualizer. Args: visualizer (Visualizer or dict, optional): A Visualizer object or a

Returns the statistics for a requested module, grouped by operator type. The operator handle determines the name associated with each

Upload a local file src to dst and return the destination file. Args: src (str or Path): A local file to be copied. d

Recursively copy an entire directory tree rooted at src to a directory named dst and return the destination directory. src and dst sh

Convert all tensors to CPU in data.

Test whether current inner iteration can be evenly divided by n. Args: batch_idx (int): Current batch index of the training, vali

Download data from ``filepath`` and write the data to local path. ``get_local_path`` is decorated by :meth:`contxtlib.contextmanager`. It

Format log string after training, validation or testing iteration. Args: runner (Runner): The runner of training phase.

Get the learning rate of the optimizer. Provide unified interface to get learning rate of optimizer. Returns: Dict[str,

Get object from name. Args: obj_name (str): The name of the object. Examples: >>> get_object_from_string('torch.optim.sgd.SG