Return DenseBlock Unit symbol for building DenseNet Parameters ---------- units_num : int the number of BasicBlock in each Den

LSTM Cell symbol

LSTM Cell symbol

Removes trailing zeros in the list of integers and returns a new list of integers

Iterates through p, identifying non-zero and non-repeating values, and returns them in a list Parameters ---------- p

Return BaiscBlock Unit symbol for building DenseBlock Parameters ---------- data : str Input data growth_rate : int

Return TransitionBlock Unit symbol for building DenseNet Parameters ---------- num_stage : int Number of stage data : str

Parameters ---------- data_root: str root directory of images data_list: str a .txt file stor

Adds Symbol.WapCTC on top of pred symbol and returns the resulting symbol

Adds Symbol.contrib.ctc_loss on top of pred symbol and returns the resulting symbol

Calculates the Longest Common Subsequence between p and l (both list of int) and returns its length

Helper function for reading in next sample.

Generate noise to a given Image based on required noise type Input parameters: image: ndarray (input image data. It will be converte

Final postprocessing step before image is loaded into the batch.

Put text, rotate, and crop. Retures an img of size IMG_WIDTH * IMG_HEIGHT. There shall be some small pixel gap (1~3) with left/top/bottom bor

Resets the iterator to the beginning of the data.

Return ResNet Unit symbol for building ResNet Parameters ---------- data : str Input data num_filter : int Number of o

Parameters ---------- data_root: str root directory of images data_list: str a .txt file stor

Simple accuracy measure: number of 100% accurate predictions divided by total number

Longest Common Subsequence accuracy measure: calculate accuracy of each prediction as LCS/length

Adds CTC loss on top of pred symbol and returns the resulting symbol

Checks if the input data is valid

Return DenseNet symbol of imagenet Parameters ---------- units : list Number of units in each stage num_stage : int