innat

import numpy as np
import tensorflow as tf
from tensorflow import keras 
from tensorflow.keras import Model, Input
from tensorflow.keras.layers import Conv2DTranspose
from tensorflow.keras.layers import UpSampling2D
from tensorflow.keras.layers import Conv2D
from tensorflow.keras.layers import BatchNormalization
from tensorflow.keras.layers import Activation
from tensorflow.keras.layers import Concatenate

innat

import tensorflow as tf
from tensorflow import keras 
from tensorflow.keras import Model 
from tensorflow.keras import layers 

class ConvoBlocks(tf.keras.layers.Layer):
    def __init__(self, num_filters=256, 
                 kernel_size=3, dilation_rate=1, 
                 padding="same", use_bias=False, **kwargs):
        super(ConvoBlocks, self).__init__(**kwargs)

innat

# Reference: https://keras.io/examples/structured_data/deep_neural_decision_forests/

import tensorflow as tf 
from tensorflow.keras import layers 
from tensorflow import keras 

class NeuralDecisionTree(keras.Model):
    def __init__(self, depth, num_features, used_features_rate, num_classes):
        super(NeuralDecisionTree, self).__init__()
        self.depth = depth

innat

import tensorflow as tf
from tensorflow.keras import layers 
from tensorflow import keras 
print('TensorFlow', tf.__version__)

class ResidualBlock(layers.Layer):
    def __init__(self, block_type=None, n_filters=None):
        super(ResidualBlock, self).__init__()
        self.n_filters = n_filters
        if block_type == 'identity':

innat

innat

Reflection Removal with Deep Learning

Few interesting model on reflection removal algorithm.

• ERRNet

• RAGNet

innat

import tensorflow as tf 
from tensorflow.keras.layers import * 
from tensorflow.keras.models import Model
from tensorflow.keras.utils import plot_model

class Conv3DBatchNorm(tf.keras.layers.Layer):
    def __init__(self, nb_filters, kernel_size, padding, strides):
        super(Conv3DBatchNorm, self).__init__()
        # parameters 
        self.nb_filters = nb_filters

innat

def vis(path1, path2, n_images, is_random=True, figsize=(16, 16)):
    '''
    https://github.com/innat
    '''
    plt.figure(figsize=figsize)
    image_names = os.listdir(path1)
    masks_names = os.listdir(path2)
    
    for i in range(n_images):
        if is_random:

innat

"""
pip install tensorflow
pip install tf2onnx keras2onnx onnxmltools
"""

import os
import pdb
import json
import traceback
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3" 

innat

import tensorflow as tf

# credit: https://stackoverflow.com/a/66524901/9215780
class CustomTrainStep(tf.keras.Model):
    def __init__(self, n_gradients, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.n_gradients = tf.constant(n_gradients, dtype=tf.int32)
        self.n_acum_step = tf.Variable(0, dtype=tf.int32, trainable=False)
        self.gradient_accumulation = [tf.Variable(tf.zeros_like(v, dtype=tf.float32), 
                                                  trainable=False) for v in self.trainable_variables]