#IMAGE FILE SIZE REDUCER PNG ELEPHANT HOW TO#

num_runs: number of baseline images to generate Returns: Averaged integrated gradients for `num_runs` baseline images ''' # 1. These steps along determine the integral approximation error. Args: img_input (ndarray): 3D image top_pred_idx: Predicted label for the input image num_steps: Number of interpolation steps between the baseline and the input used in the computation of integrated gradients. Calculate integrated gradients and return integrated_grads = ( img_input - baseline ) * avg_grads return integrated_grads def random_baseline_integrated_gradients ( img_input, top_pred_idx, num_steps = 50, num_runs = 2 ): '''Generates a number of random baseline images. Approximate the integral using the trapezoidal rule grads = ( grads + grads ) / 2.0 avg_grads = tf. expand_dims ( img, axis = 0 ) grad = get_gradients ( img, top_pred_idx = top_pred_idx ) grads. Get the gradients grads = for i, img in enumerate ( interpolated_image ): img = tf. preprocess_input ( interpolated_image ) # 3. Preprocess the interpolated images interpolated_image = xception. float32 ) interpolated_image = interpolated_image = np. Returns: Integrated gradients w.r.t input image ''' # If baseline is not provided, start with a black image # having same size as the input image. Args: img_input (ndarray): Original image top_pred_idx: Predicted label for the input image baseline (ndarray): The baseline image to start with for interpolation num_steps: Number of interpolation steps between the baseline and the input used in the computation of integrated gradients. gradient ( top_class, images ) return grads def get_integrated_gradients ( img_input, top_pred_idx, baseline = None, num_steps = 50 ): '''Computes Integrated Gradients for a predicted label. watch ( images ) preds = model ( images ) top_class = preds grads = tape. Args: img_input: 4D image tensor top_pred_idx: Predicted label for the input image Returns: Gradients of the predictions w.r.t img_input ''' images = tf. expand_dims ( array, axis = 0 ) return array def get_gradients ( img_input, top_pred_idx ): '''Computes the gradients of outputs w.r.t input image. img_to_array ( img ) # We add a dimension to transform our array into a 'batch' # of size (1, 299, 299, 3) array = np. load_img ( img_path, target_size = size ) # `array` is a float32 Numpy array of shape (299, 299, 3) array = keras. To read in-depth about integrated gradients and why this method works,ĭef get_img_array ( img_path, size = ( 299, 299 )): # `img` is a PIL image of size 299x299 img = keras. Approximate the gradients integral using the trapezoidal rule.Get the gradients for these interpolated images.Preprocess these interpolated images and do a forward pass. The authors recommend using anywhere between The steps we need in the gradient approximation for a given input image. Interpolate the baseline for a given number of steps. The shape of the baseline input needs to be A baseline input can be a black image (all pixel When making a prediction on a particular data point. Output is the last layer of our model (dense layer with softmax activation).Ĭompute which features are important to a neural network In our case, the input is an image and the To compute integrated gradients, we need to perform the following steps:

The gradient of the prediction output with regard to features of the input. Integrated Gradients is a variation on computing It isĪ model interpretability technique: you can use it to visualize the relationshipīetween input features and model predictions.

#IMAGE FILE SIZE REDUCER PNG ELEPHANT HOW TO#

Model interpretability with Integrated Gradientsĭescription: How to obtain integrated gradients for a classification model.Īttributing a classification model's prediction to its input features.

Image file size reducer png elephant