tonic.transforms
#
Module Contents#
Classes#
Composes several transforms together. This a literal copy of torchvision.transforms.Compose 

Crops events at the center to a specific output size. If output size is smaller than input 

Drops events with timestamps below min and above max. 

Drops events that are spatiotemporally not sufficiently close enough to other events in the 

Deterministically drops every nth event for every spatial location x (and potentially y). 

Randomly drops events with probability p. If random_p is selected, the drop probability is 

Drops events in a certain time interval with a length proportional to a specified ratio of 

Drops events located in a randomly chosen box area. The size of the box area is defined by a 

Drops events for individual pixels. If the locations of pixels to be dropped is known, a 

Multiplies timestamps and spatial pixel coordinates with separate factors. Useful when the 

Applies EventDrop transformation from the paper "EventDrop: Data Augmentation for Eventbased Learning". 

Applies EventDownsampling from the paper "Insectinspired Spatiotemporal Downsampling of Eventbased Input." 

Sets all polarities to zero. This transform does not have any parameters. 

Crops the sensor size to a smaller size in a random location. 

Drops all events for individual pixels with a given probability. 

Flips polarity of individual events with p. Changes polarities 1 to 0 and polarities [1, 0] 

Flips events in x with probability p. Pixels map as: 

Flips events in y with probability p. Pixels map as: 

Reverses temporal order of events with probability p. 

Sets a refractory period for each pixel, during which events will be ignored/discarded. We 

Changes x/y coordinate for each event by adding samples from a multivariate Gaussian 

Removes offset for timestamps, so that first event starts at time zero. 

Changes timestamp for each event by adding samples from a Gaussian distribution. 

Skew all event timestamps according to a linear transform. 

Adds a fixed number of n noise events that are uniformly distributed across sensor size 

Change dtype of numpy ndarray to custom dtype. This transform is necessary for example if 

Create averaged timesurfaces for each event. Taken from the paper Sironi et al. 2018, HATS: 

Accumulate events to frames by slicing along constant time (time_window), constant number of 

PyTorch sparse tensor dropin replacement for ToFrame. See 

Counts up all events to a single image of size sensor_size. 

Create global time surfaces at a specific time interval dt. 

Build a voxel grid with bilinear interpolation in the time domain from a set of events. 

Takes T*B binary event frames to produce a sequence of T frames of Nbit numbers. To do so, 

Copies target n times. 

Transforms one or more targets into a one hot encoding scheme. 
 class tonic.transforms.Compose(transforms: Callable)[source]#
Composes several transforms together. This a literal copy of torchvision.transforms.Compose function for convenience.
 Parameters:
transforms (list of
Transform
objects) – list of transform(s) to compose. Can combine Tonic, PyTorch Vision/Audio transforms.
Example
>>> transforms.Compose([ >>> transforms.Denoise(filter_time=10000), >>> transforms.ToFrame(n_time_bins=3), >>> ])
 class tonic.transforms.CenterCrop[source]#
Crops events at the center to a specific output size. If output size is smaller than input sensor size along any dimension, padding will be used, which doesn’t influence the number of events on that axis but just their spatial location after cropping. Make sure to use the cropped sensor size for any transform after CenterCrop.
 Parameters:
sensor_size (tuple) – Size of the sensor that was used [W,H,P]
size (sequence or int) – Desired output size of the crop. If size is an int instead of sequence like (h, w), a square crop (size, size) is made.
 sensor_size: Tuple[int, int, int]#
 size: Union[int, Tuple[int, int]]#
 class tonic.transforms.CropTime[source]#
Drops events with timestamps below min and above max.
 Parameters:
min (int) – The minimum timestamp below which all events are dropped. Zero by default.
max (int) – The maximum timestamp above which all events are dropped.
Example
>>> transform = tonic.transforms.CropTime(min=1000, max=20000)
 min: int = 0#
 max: int#
 class tonic.transforms.Denoise[source]#
Drops events that are spatiotemporally not sufficiently close enough to other events in the sample. In practise that means that an event is dropped if no other event occured within a spatial neighbourhood of 1 pixel and a temporal neighbourhood of filter_time time units. Useful to filter noisy recordings where events occur isolated in time.
 Parameters:
filter_time (float) – minimum temporal distance to next event, otherwise dropped. Lower values will mean higher constraints, therefore less output events. Use same unit of time as the events have.
Example
>>> transform = tonic.transforms.Denoise(filter_time=10000)
 filter_time: float#
 class tonic.transforms.Decimation[source]#
Deterministically drops every nth event for every spatial location x (and potentially y).
 Parameters:
n (int) – The event stream for each x/y location is reduced to 1/n.
Example
>>> transform = tonic.transforms.Decimation(n=5)
 n: int#
 class tonic.transforms.DropEvent[source]#
Randomly drops events with probability p. If random_p is selected, the drop probability is randomized between 0 and p.
 Parameters:
p (float or tuple of floats) – Probability of dropping events. Can be a tuple of floats (p_min, p_max), so that p is sampled from the range.
Example
>>> transform1 = tonic.transforms.DropEvent(p=0.2) >>> transform2 = tonic.transforms.DropEvent(p=(0, 0.5))
 p: Union[float, Tuple[float, float]]#
 class tonic.transforms.DropEventByTime[source]#
Drops events in a certain time interval with a length proportional to a specified ratio of the original length.
 Parameters:
duration_ratio (Union[float, Tuple[float]], optional) – the length of the dropped time interval, expressed in a ratio of the original sequence duration.  If a float, the value is used to calculate the interval length  If a tuple of 2 floats, the ratio is randomly chosen in [min, max) Defaults to 0.2.
Example
>>> transform = tonic.transforms.DropEventByTime(duration_ratio=(0.1, 0.8))
 duration_ratio: Union[float, Tuple[float, float]] = 0.2#
 class tonic.transforms.DropEventByArea[source]#
Drops events located in a randomly chosen box area. The size of the box area is defined by a specified ratio of the sensor size.
 Parameters:
sensor_size (Tuple) – size of the sensor that was used [W,H,P]
area_ratio (Union[float, Tuple[float]], optional) – Ratio of the sensor resolution that determines the size of the box area where events are dropped.  if a float, the value is used to calculate the size of the box area  if a tuple of 2 floats, the ratio is randomly chosen in [min, max) Defaults to 0.2.
Example
>>> transform = tonic.transforms.DropEventByArea(sensor_size=(128,128,2), area_ratio=(0.1, 0.8))
 sensor_size: Tuple[int, int, int]#
 area_ratio: Union[float, Tuple[float, float]] = 0.2#
 class tonic.transforms.DropPixel[source]#
Drops events for individual pixels. If the locations of pixels to be dropped is known, a list of x/y coordinates can be passed directly. Alternatively, a cutoff frequency for each pixel can be defined above which pixels will be deactivated completely. This prevents so called hot pixels which fire at a high frequency even in the absence of any input signal (e.g. due to faulty hardware).
 Parameters:
coordinates – List of (x,y) coordinates for which all events will be deleted.
hot_pixel_frequency – Drop pixels completely that fire higher than the given frequency.
Example
>>> from tonic.transforms import DropPixel >>> transform1 = DropPixel(coordinates=[[10,10], [10,11], [11,10], [11,11]]) >>> transform2 = DropPixel(hot_pixel_frequency=60) # Hertz
 coordinates: Optional[List[Tuple[int, int]]]#
 hot_pixel_frequency: Optional[int]#
 class tonic.transforms.Downsample[source]#
Multiplies timestamps and spatial pixel coordinates with separate factors. Useful when the native temporal and/or spatial resolution of the original sensor is too high for downstream processing, notably when converting to dense representations of some sort. This transform does not drop any events.
 Parameters:
time_factor (float) – value to multiply timestamps with. Default is 1.
spatial_factor (float or tuple of floats) – values to multiply pixel coordinates with. Default is 1. Note that when using subsequential transforms that require sensor_size, you must change the spatial values for the later transformation.
sensor_size (tuple) – size of the sensor that was used [W,H,P]
target_size (tuple) – size of the desired resolution [W,H]
Example
>>> from tonic.transforms import Downsample >>> transform1 = Downsample(time_factor=0.001) # change us to ms >>> transform2 = Downsample(spatial_factor=0.25) # reduce focal plane to 1/4. >>> transform3 = Downsample(sensor_size=(40, 20, 2), target_size=(10, 5)) # reduce focal plane to 1/4.
 time_factor: float = 1#
 spatial_factor: Union[float, Tuple[float, float]] = 1#
 sensor_size: Optional[Tuple[int, int, int]]#
 target_size: Optional[Tuple[int, int]]#
 class tonic.transforms.EventDrop[source]#
 Applies EventDrop transformation from the paper “EventDrop: Data Augmentation for Eventbased Learning”.
 Applies one of the 4 drops of event strategies between:
Identity (do nothing)
Drop events by time
Drop events by area
Drop events randomly
For each strategy, the ratio of dropped events are determined in the paper.
 Parameters:
sensor_size (Tuple) – size of the sensor that was used [W,H,P]
Example
>>> transform = tonic.transforms.EventDrop(sensor_size=(128,128,2))
 sensor_size: Tuple[int, int, int]#
 class tonic.transforms.EventDownsampling[source]#
 Applies EventDownsampling from the paper “Insectinspired Spatiotemporal Downsampling of Eventbased Input.”
 Allows:
Integrator based method to perform spatiotemporal eventbased downsampling
Differentiator based method to perform spatiotemporal eventbased downsampling
 Parameters:
sensor_size (Tuple) – size of the sensor that was used [W,H,P]
target_size (Tuple) – size of the desired resolution [W,H]
dt (float) – temporal resolution of events in ms
downsampling_method (str) – string stating downsampling method. Choose from [‘naive’, ‘integrator’, ‘differentiator’]
noise_threshold (int) – set number of events in downsampled pixel required to emit spike. Zero by default.
differentiator_time_bins (int) – number of differentiator time bins within dt. Two by default.
Example
>>> transform1 = tonic.transforms.EventDownsampling(sensor_size=(640,480,2), target_size=(20,15), dt=0.5, downsampling_method='integrator') >>> transform2 = tonic.transforms.EventDownsampling(sensor_size=(640,480,2), target_size=(20,15), dt=0.5, downsampling_method='differentiator', noise_threshold=2, differentiator_time_bins=3)
 sensor_size: Tuple[int, int, int]#
 target_size: Tuple[int, int]#
 downsampling_method: str#
 dt: Optional[float]#
 noise_threshold: Optional[int]#
 differentiator_time_bins: Optional[int]#
 class tonic.transforms.MergePolarities[source]#
Sets all polarities to zero. This transform does not have any parameters.
Example
>>> transform = tonic.transforms.MergePolarities()
 class tonic.transforms.RandomCrop[source]#
Crops the sensor size to a smaller size in a random location.
x’ = x  new_sensor_start_x
y’ = y  new_sensor_start_y
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
target_size – a tuple of x,y target sensor size
Example
>>> transform = tonic.transforms.RandomCrop(sensor_size=(340, 240, 2), target_size=(50, 50))
 sensor_size: Tuple[int, int, int]#
 target_size: Tuple[int, int]#
 class tonic.transforms.RandomDropPixel[source]#
Drops all events for individual pixels with a given probability.
 Parameters:
p – Probability of pixel being dropped. Stochastic transform.
sensor_size – a 3tuple of x,y,p for sensor_size. Not necessary when RandomDropPixel is applied to rasters.
Example
>>> from tonic.transforms import RandomDropPixel >>> transform = DropPixel(p=0.2)
 p: float#
 sensor_size: Optional[Tuple[int, int, int]]#
 class tonic.transforms.RandomFlipPolarity[source]#
Flips polarity of individual events with p. Changes polarities 1 to 0 and polarities [1, 0] to 1.
 Parameters:
p (float) – probability of flipping individual event polarities
Example
>>> transform = tonic.transforms.RandomFlipPolarity(p=0.3)
 p: float = 0.5#
 class tonic.transforms.RandomFlipLR[source]#
Flips events in x with probability p. Pixels map as:
x’ = width  x
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
p (float) – probability of performing the flip
Example
>>> transform = tonic.transforms.RandomFlipLR(p=0.3)
 sensor_size: Tuple[int, int, int]#
 p: float = 0.5#
 class tonic.transforms.RandomFlipUD[source]#
Flips events in y with probability p. Pixels map as:
y’ = height  y
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
p (float) – probability of performing the flip
Example
>>> transform = tonic.transforms.RandomFlipUD(p=0.3)
 sensor_size: Tuple[int, int, int]#
 p: float = 0.5#
 class tonic.transforms.RandomTimeReversal[source]#
Reverses temporal order of events with probability p.
\[t_i' = max(t)  t_i\] Parameters:
p (float) – probability of performing the flip
flip_polarities (bool) – if the time is reversed, also flip the polarities. True by default.
Example
>>> transform = tonic.transforms.RandomTimeReversal(p=0.3)
 p: float = 0.5#
 flip_polarities: bool = True#
 class tonic.transforms.RefractoryPeriod[source]#
Sets a refractory period for each pixel, during which events will be ignored/discarded. We keep events if:
\[t_n  t_{n1} > t_{refrac}\]for each pixel.
 Parameters:
delta (int) – Refractory period for each pixel. Use same time unit as event timestamps. Can use a 2tuple to sample from a range.
>>> transform1 = tonic.transforms.RefractoryPeriod(delta=1000) >>> transform2 = tonic.transforms.RefractoryPeriod(delta=[0, 1000])
 delta: Union[int, Tuple[int, int]]#
 class tonic.transforms.SpatialJitter[source]#
Changes x/y coordinate for each event by adding samples from a multivariate Gaussian distribution. It with the following properties:
\[ \begin{align}\begin{aligned}mean = [x,y]\\\Sigma = [[var_x, sigma_{xy}],[sigma_{xy}, var_y]]\end{aligned}\end{align} \]Jittered events that lie outside the focal plane will be dropped if clip_outliers is True.
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
var_x (float) – variance for the distribution in the x direction
var_y (float) – variance for the distribution in the y direction
sigma_xy (float) – changes skewness of distribution, only change if you want shifts along diagonal axis.
clip_outliers (bool) – when True, events that have been jittered outside the sensor size will be dropped.
 sensor_size: Tuple[int, int, int]#
 var_x: float = 1#
 var_y: float = 1#
 sigma_xy: float = 0#
 clip_outliers: bool = False#
 class tonic.transforms.TimeAlignment[source]#
Removes offset for timestamps, so that first event starts at time zero.
 class tonic.transforms.TimeJitter[source]#
Changes timestamp for each event by adding samples from a Gaussian distribution.
 Parameters:
std (sequence or float) – the standard deviation of the time jitter.
clip_negative (bool) – drops events that have negative timestamps.
sort_timestamps (bool) – sort the events by timestamps after jitter.
 std: float#
 clip_negative: bool = True#
 sort_timestamps: bool = False#
 class tonic.transforms.TimeSkew[source]#
Skew all event timestamps according to a linear transform.
 Parameters:
coefficient – a realvalued multiplier applied to the timestamps of the events. E.g. a coefficient of 2.0 will double the effective delay between any pair of events. Can provide a tuple for a range of values.
offset – value by which the timestamps will be shifted after multiplication by the coefficient. Negative offsets are permissible but may result in in an exception if timestamps are shifted below 0. Tuple of values might be provided as a range to sample from.
Example
>>> transform1 = TimeSkew(coefficient=1.3, offset=100) >>> transform2 = TimeSkew(coefficient=[0.8, 1.2], offset=[0, 150])
 coefficient: Union[float, Tuple[float, float]]#
 offset: Union[float, Tuple[float, float]] = 0#
 class tonic.transforms.UniformNoise[source]#
Adds a fixed number of n noise events that are uniformly distributed across sensor size dimensions such as x, y, t and p.
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
n – Number of events that are added. Can be a tuple of integers, so that n is sampled from a range.
Example
>>> transform = tonic.transforms.UniformNoise(sensor_size=(340, 240, 2), n=3000)
 sensor_size: Tuple[int, int, int]#
 n: Union[int, Tuple[int, int]]#
 class tonic.transforms.NumpyAsType[source]#
Change dtype of numpy ndarray to custom dtype. This transform is necessary for example if you want to load raw events using a PyTorch dataloader. The original events coming from any dataset in Tonic are structured numpy arrays, so that they can be indexed as events[“t”] or events[“p”] etc. Pytorch’s dataloader however does not support the conversion from structured numpy arrays to Tensors, that’s why we need to employ at least NumpyAsType(int) to convert the structured array into an unstructured one before handing it to the dataloader.
 Parameters:
dtype – data type that the array should be cast to.
Example
>>> # indexing the dataset directly provides structured numpy arrays >>> dataset = tonic.datasets.NMNIST(save_to='data') >>> events, targets = dataset[100] >>> >>> # this doesn't work >>> dataloader = torch.utils.data.DataLoader(dataset) >>> events, targets = next(iter(dataloader)) >>> >>> # we need to convert to unstructured arrays >>> transform = tonic.transforms.NumpyAsType(int) >>> dataset = tonic.datasets.NMNIST(save_to='data', transform=transform) >>> dataloader = torch.utils.data.DataLoader(dataset) >>> events, targets = next(iter(dataloader))
 dtype: numpy.dtype#
 class tonic.transforms.ToAveragedTimesurface[source]#
Create averaged timesurfaces for each event. Taken from the paper Sironi et al. 2018, HATS: Histograms of averaged time surfaces for robust eventbased object classification https://opena ccess.thecvf.com/content_cvpr_2018/papers/Sironi_HATS_Histograms_of_CVPR_2018_paper.pdf.
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
cell_size (int) – size of each square in the grid
surface_size (int) – has to be odd
time_window (float) – how far back to look for past events for the time averaging
tau (float) – time constant to decay events around occuring event with.
decay (str) – can be either ‘lin’ or ‘exp’, corresponding to linear or exponential decay.
 sensor_size: Tuple[int, int, int]#
 surface_size: int = 5#
 cell_size: int = 10#
 time_window: float = 1000.0#
 tau: float = 100#
 decay: str = 'exp'#
 class tonic.transforms.ToFrame[source]#
Accumulate events to frames by slicing along constant time (time_window), constant number of events (event_count) or constant number of frames (n_time_bins / n_event_bins). All the events in one slice are added up in a frame for each polarity. If you want binary frames, you can manually clamp them to 1 afterwards. You can set one of the first 4 parameters to choose the slicing method. Depending on which method you choose, overlap will be defined differently. As a rule of thumb, here are some considerations if you are unsure which slicing method to choose:
If your recordings are of roughly the same length, a safe option is to set time_window. Bare in mind that the number of events can vary greatly from slice to slice, but will give you some consistency when training RNNs or other algorithms that have time steps.
If your recordings have roughly the same amount of activity / number of events and you are more interested in the spatial composition, then setting event_count will give you frames that are visually more consistent.
The previous time_window and event_count methods will likely result in a different amount of frames for each recording. If your training method benefits from consistent number of frames across a dataset (for easier batching for example), or you want a parameter that is easier to set than the exact window length or number of events per slice, consider fixing the number of frames by setting n_time_bins or n_event_bins. The two methods slightly differ with respect to how the slices are distributed across the recording. You can define an overlap between 0 and 1 to provide some robustness.
 Parameters:
sensor_size – A 3tuple of x,y,p for sensor_size. If omitted, the sensor size is calculated for that sample. However, do use this feature sparingly as when not all pixels fire in a sample, this might cause issues with batching/ stacking tensors further down the line.
time_window (float) – Time window length for one frame. Use the same time unit as timestamps in the event recordings. Good if you want temporal consistency in your training, bad if you need some visual consistency for every frame if the recording’s activity is not consistent.
event_count (int) – Number of events per frame. Good for training CNNs which do not care about temporal consistency.
n_time_bins (int) – Fixed number of frames, sliced along time axis. Good for generating a predetermined number of frames which might help with batching.
n_event_bins (int) – Fixed number of frames, sliced along number of events in the recording. Good for generating a predetermined number of frames which might help with batching.
overlap (float) – Overlap between frames. The definition of overlap depends on the slicing method. For slicing by time_window, the overlap is defined in microseconds. For slicing by event_count, the overlap is defined by number of events. For slicing by n_time_bins or n_event_bins, the overlap is defined by the fraction of a bin between 0 and 1.
include_incomplete (bool) – If True, includes overhang slice when time_window or event_count is specified. Not valid for bin_count methods.
Example
>>> from tonic.transforms import ToFrame >>> transform1 = ToFrame(time_window=10000, overlap=1000, include_incomplete=True) >>> transform2 = ToFrame(event_count=3000, overlap=100, include_incomplete=True) >>> transform3 = ToFrame(n_time_bins=100, overlap=0.1)
 sensor_size: Optional[Tuple[int, int, int]]#
 time_window: Optional[float]#
 event_count: Optional[int]#
 n_time_bins: Optional[int]#
 n_event_bins: Optional[int]#
 overlap: float = 0#
 include_incomplete: bool = False#
 class tonic.transforms.ToSparseTensor[source]#
PyTorch sparse tensor dropin replacement for ToFrame. See https://pytorch.org/docs/stable/sparse.html for details about sparse tensors. The dense shape of the tensor will be (TCWH) and can be inflated by calling to_dense(). You need to have PyTorch installed for this transformation. Under the hood this transform calls ToFrame() with the same parameters, converts to a pytorch tensor and calls to_sparse().
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size. If omitted, the sensor size is calculated for that sample. However, do use this feature sparingly as when not all pixels fire in a sample, this might cause issues with batching/ stacking tensors further down the line.
time_window (float) – time window length for one frame. Use the same time unit as timestamps in the event recordings. Good if you want temporal consistency in your training, bad if you need some visual consistency for every frame if the recording’s activity is not consistent.
event_count (int) – number of events per frame. Good for training CNNs which do not care about temporal consistency.
n_time_bins (int) – fixed number of frames, sliced along time axis. Good for generating a predetermined number of frames which might help with batching.
n_event_bins (int) – fixed number of frames, sliced along number of events in the recording. Good for generating a predetermined number of frames which might help with batching.
overlap (float) – overlap between frames defined either in time units, number of events or number of bins between 0 and 1.
include_incomplete (bool) – if True, includes overhang slice when time_window or event_count is specified. Not valid for bin_count methods.
Example
>>> from tonic.transforms import ToSparseTensor >>> transform1 = ToSparseTensor(time_window=10000, overlap=300, include_incomplete=True) >>> transform2 = ToSparseTensor(event_count=3000, overlap=100, include_incomplete=True) >>> transform3 = ToSparseTensor(n_time_bins=100, overlap=0.1)
 sensor_size: Tuple[int, int, int]#
 time_window: Optional[float]#
 event_count: Optional[int]#
 n_time_bins: Optional[int]#
 n_event_bins: Optional[int]#
 overlap: float = 0#
 include_incomplete: bool = False#
 class tonic.transforms.ToImage[source]#
Counts up all events to a single image of size sensor_size.
ToImage will typically be used in combination with SlicedDataset to cut a recording into smaller chunks that are then individually binned to frames.
 sensor_size: Tuple[int, int, int]#
 class tonic.transforms.ToTimesurface[source]#
Create global time surfaces at a specific time interval dt.
 Parameters:
sensor_size – A 3tuple of x,y,p for sensor_size
dt (float) – The interval at which the timesurfaces are accumulated.
tau (float) – Time constant to decay events with.
 sensor_size: Tuple[int, int, int]#
 dt: float#
 tau: float#
 class tonic.transforms.ToVoxelGrid[source]#
Build a voxel grid with bilinear interpolation in the time domain from a set of events. Implements the event volume from Zhu et al. 2019, Unsupervised eventbased learning of optical flow, depth, and egomotion.
 Parameters:
sensor_size – a 3tuple of x,y,p for sensor_size
n_time_bins (int) – fixed number of time bins to slice the event sample into.
 sensor_size: Tuple[int, int, int]#
 n_time_bins: int#
 class tonic.transforms.ToBinaRep[source]#
Takes T*B binary event frames to produce a sequence of T frames of Nbit numbers. To do so, N binary frames are interpreted as a single frame of Nbit representation. Taken from the paper Barchid et al. 2022, BinaRep Event Frames: a Simple and Effective Representation for Event based cameras https://arxiv.org/pdf/2202.13662.pdf.
 Parameters:
n_frames (int) – the number T of binarep frames.
n_bits (int) – the number N of bits used in the Nbit representation.
Example
>>> n_time_bins = n_frames * n_bits >>> >>> transforms.Compose([ >>> transforms.ToFrame( >>> sensor_size=sensor_size, >>> n_time_bins=n_time_bins, >>> ), >>> transforms.ToBinaRep( >>> n_frames=n_frames, >>> n_bits=n_bits, >>> ), >>> ])
 n_frames: Optional[int] = 1#
 n_bits: Optional[int] = 8#