Raphael Mazzine, PhD rmazzine

🎯

Focusing

Data Scientist

rmazzine / simple_cf_gen.py

Last active July 13, 2021 11:16

Simple Counterfactual Generator

	import numpy as np
	def simple_cf_generator(factual_oh, adapted_nn, tolerance=100):
	# Get the current prediction result of the factual class
	current_pred = adapted_nn.predict(np.array([factual_oh]))[0][0]

	# The predictor will always start with a probability lower than 0.5
	if current_pred < 0.5:
	predictor = adapted_nn.predict
	else:
	predictor = lambda x: 1 - adapted_nn(x)

rmazzine / requirements.txt

Created May 8, 2022 09:38

Python Packages For DiCE loop problem