Same as add just more pythonic.

Copies new_config_dict into this config object. Note: new_config_dict can also be a config object.

Copies this config into a new config object, making the changes given by new_config_dict.

Add a log file to the list of logs being considered.

Returns the average of the elements in the window.

Add an iteration to the log with the specified data points. Type should be the type of information this is (e.g., train, valid, etc.)

Plot sequential log data.

Here one layer means a string of n blocks.

Decode predicted bbox coordinates using the same scheme employed by Yolov2: https://arxiv.org/pdf/1612.08242.pdf b_x = (sigmoid(pred

Sanitizes the input coordinates so that x1 < x2, x1 != x2, x1 >= 0, and x2 <= image_size. Also converts from relative to absolute coordinates

Here one layer means a string of n Bottleneck blocks.

Adds an element to the window, removing the earliest element if necessary.

Compute the jaccard overlap of two sets of boxes. The jaccard overlap is simply the intersection over union of two boxes. Here we operate on

Implements a conv, activation, then batch norm. Arguments are passed into the conv layer.

Adapted from https://discuss.pytorch.org/t/how-to-train-with-frozen-batchnorm/12106/8

A helper function to take a config setting and turn it into a network. Used by protonet and extrahead. Returns (network, out_channels)

Resets the MovingAverage to its initial state.

Sets the active config. Works even if cfg is already imported!

Note: config should be config.name.

Returns a dictionary of stats mined from nvidia-smi for each gpu in a list. Adapted from nvgpu: https://pypi.org/project/nvgpu/, but mine has

Postprocesses the output of Yolact on testing mode into a format that makes sense, accounting for all the possible configuration settings.

Note: If undo_transform=False then im_h and im_w are allowed to be None.

Saves the model's weights using compression because the file sizes were getting too big.