import tensorflow as tf
import numpy as np
rule_tag = tf.constant([[7],[5],[5],[5],[7]])
preds = tf.constant([[0.9, 0.1, 0.3],[1.5, 1.1, 0.2]])
print(rule_tag)
eq = tf.equal(tf.constant(7, dtype=tf.int32), rule_tag)
indices = tf.reduce_max(tf.where(eq, tf.ones_like(rule_tag), tf.zeros_like(rule_tag)), axis=1)
indices2 = tf.stack([tf.range(tf.shape(indices)[0]), tf.cast(indices, tf.int32)], axis=1)
preds_add = tf.concat([preds, tf.zeros([tf.shape(preds)[0],1])], axis=1)
sess = tf.Session()
print(preds)
print(preds_add)
print(sess.run(preds_add))
# Learns best fit is W: [0.1], b: [0.3]# your code goes here
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
Tensor("Const:0", shape=(5, 1), dtype=int32)
Tensor("Const_1:0", shape=(2, 3), dtype=float32)
Tensor("concat:0", shape=(2, 4), dtype=float32)
[[0.9 0.1 0.3 0. ]
[1.5 1.1 0.2 0. ]]