config.py

import tensorflow as tf
import tensorflow_hub as hub
from pathlib import Path
import numpy as np
import config
import os
from logger import logging


FRAME_HT = 224
FRAME_WD = 224
FRAME_NUM = 8

# tensorflow urls to download the model and lables
TENSORFLOW_HUB_URL_LABELS = "https://raw.githubusercontent.com/tensorflow/models/f8af2291cced43fc9f1d9b41ddbf772ae7b0d7d2/official/projects/movinet/files/kinetics_600_labels.txt"
TENSORFLOW_HUB_URL_MODEL = "https://tfhub.dev/tensorflow/movinet/a2/base/kinetics-600/classification/3"

MODEL_PATH = os.path.join(os.getcwd(), 'models', 'Activity_recognition.h5')


def get_labels() :
 """
 Downloads and saves the labels for tensorflow 'movienet' model.

 Returns the path of the file 'labels.txt' where the labels are saved.
 """

 logging.info(">>> Downloading the labels 'movienet' model... ")

 labels_path = tf.keras.utils.get_file(
 fname=os.path.join(os.getcwd(), 'labels.txt'),
 origin=config.TENSORFLOW_HUB_URL_LABELS
 )
 
 labels_path = Path(labels_path)

 lines = labels_path.read_text().splitlines()
 KINETICS_600_LABELS = np.array([line.strip() for line in lines])

 logging.info("Labels retrieved successfully.")

 return KINETICS_600_LABELS


def get_model() -> tf.keras.models.Model :
 """
 Downloads the tensorflow 'movienet' model. 

 Returns tensorflow.keras.models.Model object instance.
 """

 logging.info(">>> Downloading the 'movienet' model from tensorflow...")

 encoder = hub.KerasLayer(TENSORFLOW_HUB_URL_MODEL, trainable=True)

 inputs = tf.keras.layers.Input(
 shape=[FRAME_NUM, FRAME_HT, FRAME_WD, 3],
 dtype=tf.float32,
 name='image'
 )

 outputs = encoder(dict(image=inputs))

 model = tf.keras.Model(inputs, outputs, name='movinet')

 logging.info("Model downloaded successfully.")

 return model


MODEL = get_model()
KINETICS_600_LABELS = get_labels()