Week 2. Why we use K.mean here?

OdessitUA · April 22, 2023, 1:51pm

I don’t understand why we use K.mean in Contrastive Loss in class and don’t use in other places?
This is screen from video “Coding Contrastive Loss”

ai_curious · April 22, 2023, 3:32pm

It’s an interesting question. Notice that averaging is not being done in this fragment from the TensorFlow implementation…

return y_true * tf.math.square(y_pred) + (1.0 - y_true) * tf.math.square(
tf.math.maximum(margin - y_pred, 0.0)

github.com

tensorflow/addons/blob/v0.17.0/tensorflow_addons/losses/contrastive.py

# Copyright 2019 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Implements contrastive loss."""

import tensorflow as tf
from typeguard import typechecked

from tensorflow_addons.utils.keras_utils import LossFunctionWrapper

This file has been truncated. show original

It isn’t mentioned in the video?

OdessitUA · April 22, 2023, 5:56pm

The teacher skipped this line.

gent.spah · April 23, 2023, 11:37am

I did some research on this and it seems the instructor has mistaken that calculation with the RMSE calculation, seems to me. I will raise an issue in github so they can have a look at it as well.