Public Member Functions
	__init__ (self, start_iter=0)

	put_image (self, img_name, img_tensor)

	put_scalar (self, name, value, smoothing_hint=True)

	put_scalars (self, smoothing_hint=True, *kwargs)

	put_histogram (self, hist_name, hist_tensor, bins=1000)

	history (self, name)

	histories (self)

	latest (self)

	latest_with_smoothing_hint (self, window_size=20)

	smoothing_hints (self)

	step (self)

	iter (self)

	iteration (self)

	__enter__ (self)

	__exit__ (self, exc_type, exc_val, exc_tb)

	name_scope (self, name)

	clear_images (self)

	clear_histograms (self)

Protected Attributes
	_history

	_smoothing_hints

	_latest_scalars

	_iter

	_current_prefix

	_vis_data

	_histograms

Detailed Description

The user-facing class that provides metric storage functionalities.
In the future we may add support for storing / logging other types of data if needed.

Definition at line 250 of file events.py.

Constructor & Destructor Documentation

◆ init()

fastreid.utils.events.EventStorage.__init__	(		self,
			start_iter = 0 )

Args:
    start_iter (int): the iteration number to start with

Definition at line 256 of file events.py.

    def __init__(self, start_iter=0):
        """
        Args:
            start_iter (int): the iteration number to start with
        """
        self._history = defaultdict(HistoryBuffer)
        self._smoothing_hints = {}
        self._latest_scalars = {}
        self._iter = start_iter
        self._current_prefix = ""
        self._vis_data = []
        self._histograms = []
 

Member Function Documentation

◆ enter()

fastreid.utils.events.EventStorage.__enter__ ( self )

Definition at line 413 of file events.py.

    def __enter__(self):
        _CURRENT_STORAGE_STACK.append(self)
        return self
 

◆ exit()

fastreid.utils.events.EventStorage.__exit__	(	self,
		exc_type,
		exc_val,
		exc_tb )

Definition at line 417 of file events.py.

    def __exit__(self, exc_type, exc_val, exc_tb):
        assert _CURRENT_STORAGE_STACK[-1] == self
        _CURRENT_STORAGE_STACK.pop()
 

◆ clear_histograms()

fastreid.utils.events.EventStorage.clear_histograms ( self )

Delete all the stored histograms for visualization.
This should be called after histograms are written to tensorboard.

Definition at line 440 of file events.py.

    def clear_histograms(self):
        """
        Delete all the stored histograms for visualization.
        This should be called after histograms are written to tensorboard.
        """
        self._histograms = []

◆ clear_images()

fastreid.utils.events.EventStorage.clear_images ( self )

Delete all the stored images for visualization. This should be called
after images are written to tensorboard.

Definition at line 433 of file events.py.

    def clear_images(self):
        """
        Delete all the stored images for visualization. This should be called
        after images are written to tensorboard.
        """
        self._vis_data = []
 

◆ histories()

fastreid.utils.events.EventStorage.histories ( self )

Returns:
    dict[name -> HistoryBuffer]: the HistoryBuffer for all scalars

Definition at line 356 of file events.py.

    def histories(self):
        """
        Returns:
            dict[name -> HistoryBuffer]: the HistoryBuffer for all scalars
        """
        return self._history
 

◆ history()

fastreid.utils.events.EventStorage.history	(		self,
			name )

Returns:
    HistoryBuffer: the scalar history for name

Definition at line 346 of file events.py.

    def history(self, name):
        """
        Returns:
            HistoryBuffer: the scalar history for name
        """
        ret = self._history.get(name, None)
        if ret is None:
            raise KeyError("No history metric available for {}!".format(name))
        return ret
 

◆ iter()

fastreid.utils.events.EventStorage.iter ( self )

Definition at line 405 of file events.py.

    def iter(self):
        return self._iter
 

◆ iteration()

fastreid.utils.events.EventStorage.iteration ( self )

Definition at line 409 of file events.py.

    def iteration(self):
        # for backward compatibility
        return self._iter
 

◆ latest()

fastreid.utils.events.EventStorage.latest ( self )

Returns:
    dict[str -> (float, int)]: mapping from the name of each scalar to the most
        recent value and the iteration number its added.

Definition at line 363 of file events.py.

    def latest(self):
        """
        Returns:
            dict[str -> (float, int)]: mapping from the name of each scalar to the most
                recent value and the iteration number its added.
        """
        return self._latest_scalars
 

◆ latest_with_smoothing_hint()

fastreid.utils.events.EventStorage.latest_with_smoothing_hint	(		self,
			window_size = 20 )

Similar to :meth:`latest`, but the returned values
are either the un-smoothed original latest value,
or a median of the given window_size,
depend on whether the smoothing_hint is True.
This provides a default behavior that other writers can use.

Definition at line 371 of file events.py.

    def latest_with_smoothing_hint(self, window_size=20):
        """
        Similar to :meth:`latest`, but the returned values
        are either the un-smoothed original latest value,
        or a median of the given window_size,
        depend on whether the smoothing_hint is True.
        This provides a default behavior that other writers can use.
        """
        result = {}
        for k, (v, itr) in self._latest_scalars.items():
            result[k] = (
                self._history[k].median(window_size) if self._smoothing_hints[k] else v,
                itr,
            )
        return result
 

◆ name_scope()

fastreid.utils.events.EventStorage.name_scope	(		self,
			name )

Yields:
    A context within which all the events added to this storage
    will be prefixed by the name scope.

Definition at line 422 of file events.py.

    def name_scope(self, name):
        """
        Yields:
            A context within which all the events added to this storage
            will be prefixed by the name scope.
        """
        old_prefix = self._current_prefix
        self._current_prefix = name.rstrip("/") + "/"
        yield
        self._current_prefix = old_prefix
 

◆ put_histogram()

fastreid.utils.events.EventStorage.put_histogram	(	self,
		hist_name,
		hist_tensor,
		bins = 1000 )

Create a histogram from a tensor.
Args:
    hist_name (str): The name of the histogram to put into tensorboard.
    hist_tensor (torch.Tensor): A Tensor of arbitrary shape to be converted
        into a histogram.
    bins (int): Number of histogram bins.

Definition at line 317 of file events.py.

    def put_histogram(self, hist_name, hist_tensor, bins=1000):
        """
        Create a histogram from a tensor.
        Args:
            hist_name (str): The name of the histogram to put into tensorboard.
            hist_tensor (torch.Tensor): A Tensor of arbitrary shape to be converted
                into a histogram.
            bins (int): Number of histogram bins.
        """
        ht_min, ht_max = hist_tensor.min().item(), hist_tensor.max().item()
 
        # Create a histogram with PyTorch
        hist_counts = torch.histc(hist_tensor, bins=bins)
        hist_edges = torch.linspace(start=ht_min, end=ht_max, steps=bins + 1, dtype=torch.float32)
 
        # Parameter for the add_histogram_raw function of SummaryWriter
        hist_params = dict(
            tag=hist_name,
            min=ht_min,
            max=ht_max,
            num=len(hist_tensor),
            sum=float(hist_tensor.sum()),
            sum_squares=float(torch.sum(hist_tensor ** 2)),
            bucket_limits=hist_edges[1:].tolist(),
            bucket_counts=hist_counts.tolist(),
            global_step=self._iter,
        )
        self._histograms.append(hist_params)
 

◆ put_image()

fastreid.utils.events.EventStorage.put_image	(	self,
		img_name,
		img_tensor )

Add an `img_tensor` associated with `img_name`, to be shown on
tensorboard.
Args:
    img_name (str): The name of the image to put into tensorboard.
    img_tensor (torch.Tensor or numpy.array): An `uint8` or `float`
        Tensor of shape `[channel, height, width]` where `channel` is
        3. The image format should be RGB. The elements in img_tensor
        can either have values in [0, 1] (float32) or [0, 255] (uint8).
        The `img_tensor` will be visualized in tensorboard.

Definition at line 269 of file events.py.

    def put_image(self, img_name, img_tensor):
        """
        Add an `img_tensor` associated with `img_name`, to be shown on
        tensorboard.
        Args:
            img_name (str): The name of the image to put into tensorboard.
            img_tensor (torch.Tensor or numpy.array): An `uint8` or `float`
                Tensor of shape `[channel, height, width]` where `channel` is
                3. The image format should be RGB. The elements in img_tensor
                can either have values in [0, 1] (float32) or [0, 255] (uint8).
                The `img_tensor` will be visualized in tensorboard.
        """
        self._vis_data.append((img_name, img_tensor, self._iter))
 

◆ put_scalar()

fastreid.utils.events.EventStorage.put_scalar	(	self,
		name,
		value,
		smoothing_hint = True )

Add a scalar `value` to the `HistoryBuffer` associated with `name`.
Args:
    smoothing_hint (bool): a 'hint' on whether this scalar is noisy and should be
        smoothed when logged. The hint will be accessible through
        :meth:`EventStorage.smoothing_hints`.  A writer may ignore the hint
        and apply custom smoothing rule.
        It defaults to True because most scalars we save need to be smoothed to
        provide any useful signal.

Definition at line 283 of file events.py.

    def put_scalar(self, name, value, smoothing_hint=True):
        """
        Add a scalar `value` to the `HistoryBuffer` associated with `name`.
        Args:
            smoothing_hint (bool): a 'hint' on whether this scalar is noisy and should be
                smoothed when logged. The hint will be accessible through
                :meth:`EventStorage.smoothing_hints`.  A writer may ignore the hint
                and apply custom smoothing rule.
                It defaults to True because most scalars we save need to be smoothed to
                provide any useful signal.
        """
        name = self._current_prefix + name
        history = self._history[name]
        value = float(value)
        history.update(value, self._iter)
        self._latest_scalars[name] = (value, self._iter)
 
        existing_hint = self._smoothing_hints.get(name)
        if existing_hint is not None:
            assert (
                existing_hint == smoothing_hint
            ), "Scalar {} was put with a different smoothing_hint!".format(name)
        else:
            self._smoothing_hints[name] = smoothing_hint
 

◆ put_scalars()

fastreid.utils.events.EventStorage.put_scalars	(		self,
		*	smoothing_hint = True,
		**	kwargs )

Put multiple scalars from keyword arguments.
Examples:
    storage.put_scalars(loss=my_loss, accuracy=my_accuracy, smoothing_hint=True)

Definition at line 308 of file events.py.

    def put_scalars(self, *, smoothing_hint=True, **kwargs):
        """
        Put multiple scalars from keyword arguments.
        Examples:
            storage.put_scalars(loss=my_loss, accuracy=my_accuracy, smoothing_hint=True)
        """
        for k, v in kwargs.items():
            self.put_scalar(k, v, smoothing_hint=smoothing_hint)
 

◆ smoothing_hints()

fastreid.utils.events.EventStorage.smoothing_hints ( self )

Returns:
    dict[name -> bool]: the user-provided hint on whether the scalar
        is noisy and needs smoothing.

Definition at line 387 of file events.py.

    def smoothing_hints(self):
        """
        Returns:
            dict[name -> bool]: the user-provided hint on whether the scalar
                is noisy and needs smoothing.
        """
        return self._smoothing_hints
 

◆ step()

fastreid.utils.events.EventStorage.step ( self )

User should call this function at the beginning of each iteration, to
notify the storage of the start of a new iteration.
The storage will then be able to associate the new data with the
correct iteration number.

Definition at line 395 of file events.py.

    def step(self):
        """
        User should call this function at the beginning of each iteration, to
        notify the storage of the start of a new iteration.
        The storage will then be able to associate the new data with the
        correct iteration number.
        """
        self._iter += 1
 

Member Data Documentation

◆ _current_prefix

fastreid.utils.events.EventStorage._current_prefix

protected

Definition at line 265 of file events.py.

◆ _histograms

fastreid.utils.events.EventStorage._histograms

protected

Definition at line 267 of file events.py.

◆ _history

fastreid.utils.events.EventStorage._history

protected

Definition at line 261 of file events.py.

◆ _iter

fastreid.utils.events.EventStorage._iter

protected

Definition at line 264 of file events.py.

◆ _latest_scalars

fastreid.utils.events.EventStorage._latest_scalars

protected

Definition at line 263 of file events.py.

◆ _smoothing_hints

fastreid.utils.events.EventStorage._smoothing_hints

protected

Definition at line 262 of file events.py.

◆ _vis_data

fastreid.utils.events.EventStorage._vis_data

protected

Definition at line 266 of file events.py.

The documentation for this class was generated from the following file:

smreid/fastreid/utils/events.py

Public Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ __init__()

Member Function Documentation

◆ __enter__()

◆ __exit__()

◆ clear_histograms()

◆ clear_images()

◆ histories()

◆ history()

◆ iter()

◆ iteration()

◆ latest()

◆ latest_with_smoothing_hint()

◆ name_scope()

◆ put_histogram()

◆ put_image()

◆ put_scalar()

◆ put_scalars()

◆ smoothing_hints()

◆ step()

Member Data Documentation

◆ _current_prefix

◆ _histograms

◆ _history

◆ _iter

◆ _latest_scalars

◆ _smoothing_hints

◆ _vis_data

◆ init()

◆ enter()

◆ exit()