ptmodels 0.1.3

A image classifier using pre-trained models

ptmodels

ptmodels uses pre-trained models to evaluate image datasets and helps to understand which model works better.ptmodels.

Table-of-Contents • Installation • How To Use • Classification • Credits • Related • License

Table of Contents

• Installation
  • Stable Release
  • From Sources
• How to Use
  • Preprocessing
    • Import Library
    • PreTrainedModels Class
    • Available Models
    • Load Dataset
    • load_models() Method
    • models_name() Method
  • Classification
    • fit() Method
    • train_specific_model() Method
  • Evaluation
    • evaluate_saved_model() Method
    • predict_image_saved_model() Method
• Credits
• Related
• Support
• You May Also Like
• License

Why must you document your project?

Installation

Stable Release

• Install ptmodels by using pip install into the terminal

pip install ptmodels

From Sources

• To install a Python package from a source other than the PyPI repository, you can download and unpack the source distribution yourself, and then use its setup.py script to install the package in the user site-packages directory:
• Set up your user environment
• Use the wget command to download the distribution archive (for example, 0.1.0.tar.gz) from the source (for example, https://github.com/rafsunsheikh/ptmodels/archive/refs/tags/0.1.0.tar.gz); for example:

wget https://github.com/rafsunsheikh/ptmodels/archive/refs/tags/0.1.0.tar.gz

• Use tar to unpack the archive. For example:

tar ptmodels-0.1.0.tar.gz

The distribution should unpack into a similarly-named directory in your home directory (for example, ~/ptmodels-0.1.0).

• Change (cd) to the new directory, and then, on the command line, enter:

python setup.py install --user

The --user option directs setup.py to install the package (for example, ptmodels) in the user site-packages directory for the running Python; for example:

~/.local/lib/pythonX.Y/site-packages/ptmodels

Python automatically searches this directory for modules, so prepending this path to the PYTHONPATH environmental variable is not necessary.

If you omit the --user option, setup.py will try to install the package in the global site-packages directory (where you do not have the necessary permissions); as a result, the installation will fail.

Alternatively, you can use the --home or --prefix option to install your package in a different location (where you have the necessary permissions); for example, to install your package in a subdirectory (for example, python-pkgs):

• Within your home directory, enter:

python setup.py install --home=~/python-pkgs

How To Use

Preprocessing

• Open up Kaggle Notebook or Google Colab.
• Choose runtime with GPU support. (as the models are very large, it will be difficult to run it on your local machine and also without GPU.)
• Install ptmodels in one cell.

!pip install ptmodels

• Run the cell by pressing shift + enter

Import Library

Import PreTrainedModels class from ptmodels.Classifier package.

from ptmodels.Classifier import PreTrainedModels

PreTrainedModels Class

• PreTrainedModels class is the only class of this package. It's an easy-to-use class with some parameters. You can call the PreTrainedModels class without any arguments. That time default values will be set in the parametes. Or else you can provide your own arguments values.

Arguments

• NUM_CLASSES : Takes the number of classes for classification. Default value is set to 2.
• BATCH_SIZE : Trains the image dataset into batches. Defaults Batch size value is 32.
• EPOCHS : Takes the number of number of epochs for the training. Default value set to 1.
• LEARNING_RATE : Sets the learning rate for training the model. Default value set to 1e-4.
• MOMENTUM : Sets the value of momentum for training the model. Default value set to 0.9.

Example

model = PreTrainedModels(NUM_CLASSES=10, BATCH_SIZE=32, EPOCHS=1, LEARNING_RATE=0.001, MOMENTUM=0.9)

• You can also consuly my Google Colab Notebook

Available Models

Following are the available models for ptmodels package:

Model	Size (MB)	Top-1 Accuracy	Top-5 Accuracy	Parameters	Depth	Time (ms) per inference step (CPU)	Time (ms) per inference step (GPU)
VGG16	528	71.3%	90.1%	138.4M	16	69.5	4.2
VGG19	549	71.3%	90.0%	143.7M	19	84.8	4.4
ResNet50	98	74.9%	92.1%	25.6M	107	58.2	4.6
ResNet50V2	98	76.0%	93.0%	25.6M	103	45.6	4.4
ResNet101	171	76.4%	92.8%	44.7M	209	89.6	5.2
ResNet101V2	171	77.2%	93.8%	44.7M	205	72.7	5.4
ResNet152	232	76.6%	93.1%	60.4M	311	127.4	6.5
ResNet152V2	232	78.0%	94.2%	60.4M	307	107.5	6.6
MobileNet	16	70.4%	89.5%	4.3M	55	22.6	3.4
MobileNetV2	14	71.3%	90.1%	3.5M	105	25.9	3.8
DenseNet121	33	75.0%	92.3%	8.1M	242	77.1	5.4
DenseNet169	57	76.2%	93.2%	14.3M	338	96.4	6.3
DenseNet201	80	77.3%	93.6%	20.2M	402	127.2	6.7
NASNetMobile	23	74.4%	91.9%	5.3M	389	27.0	6.7
NASNetLarge	343	82.5%	96.0%	88.9M	533	344.5	20.0
EfficientNetB0	29	77.1%	93.3%	5.3M	132	46.0	4.9
EfficientNetB1	31	79.1%	94.4%	7.9M	186	60.2	5.6
EfficientNetB2	36	80.1%	94.9%	9.2M	186	80.8	6.5
EfficientNetB3	48	81.6%	95.7%	12.3M	210	140.0	8.8
EfficientNetB4	75	82.9%	96.4%	19.5M	258	308.3	15.1
EfficientNetB5	118	83.6%	96.7%	30.6M	312	579.2	25.3
EfficientNetB6	166	84.0%	96.8%	43.3M	360	958.1	40.4
EfficientNetB7	256	84.3%	97.0%	66.7M	438	1578.9	61.6
EfficientNetV2B0	29	78.7%	94.3%	7.2M	-	-	-
EfficientNetV2B1	34	79.8%	95.0%	8.2M	-	-	-
EfficientNetV2B2	42	80.5%	95.1%	10.2M	-	-	-
EfficientNetV2B3	59	82.0%	95.8%	14.5M	-	-	-
EfficientNetV2S	88	83.9%	96.7%	21.6M	-	-	-
EfficientNetV2M	220	85.3%	97.4%	54.4M	-	-	-
EfficientNetV2L	479	85.7%	97.5%	119.0M	-	-	-
ConvNeXtTiny	109.42	81.3%	-	28.6M	-	-	-
ConvNeXtSmall	192.29	82.3%	-	50.2M	-	-	-
ConvNeXtBase	338.58	85.3%	-	88.5M	-	-	-
ConvNeXtLarge	755.07	86.3%	-	197.7M	-	-	-
ConvNeXtXLarge	1310	86.7%	-	350.1M	-	-	-

• All these models are taken from Keras Application API. These models are used as image classification using pre-trained weights. Prediction, feature extraction, and fine-tuning can be done using these models.

Load Dataset

• You can use any image dataset available online or use your own image dataset.
• Load your dataset in the preffered way.
• For this documentation we have used Cifar10 image dataset that is freely available in the keras.datasets.
• This dataset has 50,000 32x32 colored training images and 10,000 colored test images. There are 10 categories in this dataset. See more info at the CIFAR homepage.

Example:

• We are getting the dataset from tensorflow.keras.datasets.
• You can get your own dataset from anywhere.

from tensorflow.keras.datasets import cifar10
(x_train, y_train), (x_test, y_test) = cifar10.load_data()

• After loading the dataset please split the dataset into (x_train, y_train), (x_test, y_test).
• For that you can use sklearn.model_selection.train_test_split(). For that you can follow the guidelines from sklearn train_test_split()

load_models() Method

• Load your models by using load_models() method from ptmodels package.

Arguments:

• x_train : provide your x_train from loaded dataset. It is required to get the width, height and channel of the images.

Example :

model.load_models(x_train)

• This method is going to take a while for lading all the pre-trained models and their weights. You are going to see the following into your terminal.

Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5
58889256/58889256 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5
80134624/80134624 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5
94765736/94765736 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50v2_weights_tf_dim_ordering_tf_kernels_notop.h5
94668760/94668760 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet101_weights_tf_dim_ordering_tf_kernels_notop.h5
171446536/171446536 [==============================] - 2s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet101v2_weights_tf_dim_ordering_tf_kernels_notop.h5
171317808/171317808 [==============================] - 2s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet152_weights_tf_dim_ordering_tf_kernels_notop.h5
234698864/234698864 [==============================] - 2s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet152v2_weights_tf_dim_ordering_tf_kernels_notop.h5
234545216/234545216 [==============================] - 2s 0us/step
WARNING:tensorflow:`input_shape` is undefined or non-square, or `rows` is not in [128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/mobilenet/mobilenet_1_0_224_tf_no_top.h5
17225924/17225924 [==============================] - 0s 0us/step
WARNING:tensorflow:`input_shape` is undefined or non-square, or `rows` is not in [96, 128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/mobilenet_v2/mobilenet_v2_weights_tf_dim_ordering_tf_kernels_1.0_224_no_top.h5
9406464/9406464 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/densenet/densenet121_weights_tf_dim_ordering_tf_kernels_notop.h5
29084464/29084464 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/densenet/densenet169_weights_tf_dim_ordering_tf_kernels_notop.h5
51877672/51877672 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/densenet/densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5
74836368/74836368 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/nasnet/NASNet-mobile-no-top.h5
19993432/19993432 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/nasnet/NASNet-large-no-top.h5
343610240/343610240 [==============================] - 3s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb0_notop.h5
16705208/16705208 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb1_notop.h5
27018416/27018416 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb2_notop.h5
31790344/31790344 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb3_notop.h5
43941136/43941136 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb4_notop.h5
71686520/71686520 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb5_notop.h5
115263384/115263384 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb6_notop.h5
165234480/165234480 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb7_notop.h5
258076736/258076736 [==============================] - 2s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-b0_notop.h5
24274472/24274472 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-b1_notop.h5
28456008/28456008 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-b2_notop.h5
35839040/35839040 [==============================] - 0s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-b3_notop.h5
52606240/52606240 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-s_notop.h5
82420632/82420632 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-m_notop.h5
214201816/214201816 [==============================] - 2s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/efficientnet_v2/efficientnetv2-l_notop.h5
473176280/473176280 [==============================] - 5s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/convnext/convnext_tiny_notop.h5
111650432/111650432 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/convnext/convnext_small_notop.h5
198551472/198551472 [==============================] - 1s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/convnext/convnext_base_notop.h5
350926856/350926856 [==============================] - 3s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/convnext/convnext_large_notop.h5
785596384/785596384 [==============================] - 7s 0us/step
Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/convnext/convnext_xlarge_notop.h5
 853016576/1393257616 [=================>............] - ETA: 5s

models_name() Method

• models_name() method will let you find out the loaded models names in your class object.
• No arguments are needed for this method.

Example:

names = []
names = model.models_name()
print(names)

Output:

['VGG16', 'VGG19', 'ResNet50', 'ResNet50V2', 'ResNet101', 'ResNet101V2', 'ResNet152', 'ResNet152V2', 'MobileNet', 'MobileNetV2', 'DenseNet121', 'DenseNet169', 'EfficientNetV2B1', 'EfficientNetV2B2', 'EfficientNetV2B3', 'EfficientNetV2S', 'EfficientNetV2M', 'EfficientNetV2L', 'ConvNeXtTiny', 'ConvNeXtSmall', 'ConvNeXtBase', 'ConvNeXtLarge', 'ConvNeXtXLarge']

Classification

• Using ptmodels package you can easily train your dataset using all the pre-trained models or using any specific models. All the small tasks are done already. You just need to provide the dataset to train using this package.
• For image classification, we are going to use two methods.
  • fit() Method
  • train_specific_model() Method

fit() Method

• Using this method you can very easily train your dataset using all the pre-trained models.
• That will take a lot of time.
• So our suggestion is that, use a high GPU PC or use Google Colab Notebook or Kaggle Notebook and load GPU into runtime.

Arguments:

• x_train : x_train lists the images for train datasets without labels.
• y_train : y_train holds the labels for train datasets.
• x_test : x_test has the images for test dataset without labels.
• y_test : y_test has the labels for the test dataset.

Returns:

• pandas.dataframe which holds the evaluation metrics of all the pre-trained models.

Example :

dataframe = model.fit(x_train, y_train, x_test, y_test)

• This method also saves your return dataframe into prediction.csv file in your disk.

Output:

Starting Training: VGG16
1563/1563 [==============================] - 24s 10ms/step - loss: 0.2274 - accuracy: 0.5063
1563/1563 [==============================] - 11s 7ms/step
313/313 [==============================] - 2s 7ms/step
appended successfully

Starting Training: VGG19
1563/1563 [==============================] - 19s 11ms/step - loss: 0.2313 - accuracy: 0.4969
1563/1563 [==============================] - 14s 9ms/step
313/313 [==============================] - 3s 9ms/step
appended successfully

Starting Training: ResNet50
1563/1563 [==============================] - 27s 14ms/step - loss: 0.2331 - accuracy: 0.4970
1563/1563 [==============================] - 18s 10ms/step
313/313 [==============================] - 4s 12ms/step
appended successfully

Starting Training: ResNet50V2
1563/1563 [==============================] - 26s 14ms/step - loss: 0.2843 - accuracy: 0.3291
1563/1563 [==============================] - 17s 10ms/step
313/313 [==============================] - 3s 10ms/step
appended successfully

Starting Training: ResNet101
1563/1563 [==============================] - 43s 23ms/step - loss: 0.2363 - accuracy: 0.4896
1563/1563 [==============================] - 31s 18ms/step
313/313 [==============================] - 6s 20ms/step
appended successfully

Starting Training: ResNet101V2
1563/1563 [==============================] - 43s 22ms/step - loss: 0.2923 - accuracy: 0.3033
1563/1563 [==============================] - 30s 18ms/step
313/313 [==============================] - 6s 19ms/step
appended successfully

Starting Training: ResNet152
1563/1563 [==============================] - 61s 32ms/step - loss: 0.2327 - accuracy: 0.5006
1563/1563 [==============================] - 44s 26ms/step
313/313 [==============================] - 9s 28ms/step
appended successfully

Starting Training: ResNet152V2
1563/1563 [==============================] - 62s 33ms/step - loss: 0.2959 - accuracy: 0.3003
1563/1563 [==============================] - 44s 26ms/step
313/313 [==============================] - 9s 27ms/step
appended successfully

Starting Training: MobileNet
1563/1563 [==============================] - 14s 8ms/step - loss: 0.3206 - accuracy: 0.1784
1563/1563 [==============================] - 8s 5ms/step
313/313 [==============================] - 2s 5ms/step
appended successfully

Starting Training: MobileNetV2
1563/1563 [==============================] - 20s 11ms/step - loss: 0.3165 - accuracy: 0.1961
1563/1563 [==============================] - 13s 8ms/step
313/313 [==============================] - 2s 7ms/step
appended successfully

Starting Training: DenseNet121
1563/1563 [==============================] - 42s 21ms/step - loss: 0.2424 - accuracy: 0.4628
1563/1563 [==============================] - 30s 18ms/step
313/313 [==============================] - 6s 19ms/step
appended successfully

Starting Training: DenseNet169
1563/1563 [==============================] - 55s 28ms/step - loss: 0.2432 - accuracy: 0.4567
1563/1563 [==============================] - 41s 24ms/step
313/313 [==============================] - 7s 23ms/step
appended successfully

Starting Training: EfficientNetV2B1
1563/1563 [==============================] - 33s 14ms/step - loss: 0.2233 - accuracy: 0.5198
1563/1563 [==============================] - 28s 17ms/step
313/313 [==============================] - 5s 15ms/step
appended successfully

Starting Training: EfficientNetV2B2
1563/1563 [==============================] - 35s 15ms/step - loss: 0.2279 - accuracy: 0.5117
1563/1563 [==============================] - 29s 17ms/step
313/313 [==============================] - 5s 16ms/step
appended successfully

Starting Training: EfficientNetV2B3
1563/1563 [==============================] - 39s 17ms/step - loss: 0.2584 - accuracy: 0.4208
1563/1563 [==============================] - 34s 20ms/step
313/313 [==============================] - 6s 19ms/step
appended successfully

Starting Training: EfficientNetV2S
1563/1563 [==============================] - 47s 20ms/step - loss: 0.2455 - accuracy: 0.4598
1563/1563 [==============================] - 43s 25ms/step
313/313 [==============================] - 7s 24ms/step
appended successfully

Starting Training: EfficientNetV2M
1563/1563 [==============================] - 66s 28ms/step - loss: 0.3221 - accuracy: 0.2315
1563/1563 [==============================] - 60s 35ms/step
313/313 [==============================] - 11s 35ms/step
appended successfully

Starting Training: EfficientNetV2L
1563/1563 [==============================] - 95s 40ms/step - loss: 0.2644 - accuracy: 0.4007
1563/1563 [==============================] - 83s 48ms/step
313/313 [==============================] - 15s 48ms/step
appended successfully

Starting Training: ConvNeXtTiny
1563/1563 [==============================] - 60s 30ms/step - loss: 0.1644 - accuracy: 0.6741
1563/1563 [==============================] - 45s 27ms/step
313/313 [==============================] - 9s 27ms/step
appended successfully

Starting Training: ConvNeXtSmall
1563/1563 [==============================] - 103s 53ms/step - loss: 0.1573 - accuracy: 0.6885
1563/1563 [==============================] - 83s 49ms/step
313/313 [==============================] - 16s 50ms/step
appended successfully

Starting Training: ConvNeXtBase
1563/1563 [==============================] - 105s 54ms/step - loss: 0.1435 - accuracy: 0.7221
1563/1563 [==============================] - 83s 50ms/step
313/313 [==============================] - 15s 48ms/step
appended successfully

Starting Training: ConvNeXtLarge
1563/1563 [==============================] - 117s 62ms/step - loss: 0.1237 - accuracy: 0.7662
1563/1563 [==============================] - 91s 55ms/step
313/313 [==============================] - 18s 58ms/step
appended successfully

Starting Training: ConvNeXtXLarge
1563/1563 [==============================] - 129s 69ms/step - loss: 0.1178 - accuracy: 0.7845
1563/1563 [==============================] - 104s 63ms/step
313/313 [==============================] - 20s 64ms/step
appended successfully

              Models Accuracy train Precision train Recall train  \
0              VGG16          0.621           0.621        0.621   
1              VGG19          0.602           0.602        0.602   
2           ResNet50          0.622           0.622        0.622   
3         ResNet50V2          0.412           0.412        0.412   
4          ResNet101          0.614           0.614        0.614   
5        ResNet101V2          0.386           0.386        0.386   
6          ResNet152          0.622           0.622        0.622   
7        ResNet152V2          0.379           0.379        0.379   
8          MobileNet          0.205           0.205        0.205   
9        MobileNetV2          0.234           0.234        0.234   
10       DenseNet121          0.578           0.578        0.578   
11       DenseNet169          0.557           0.557        0.557   
12  EfficientNetV2B1          0.652           0.652        0.652   
13  EfficientNetV2B2          0.641           0.641        0.641   
14  EfficientNetV2B3          0.593           0.593        0.593   
15   EfficientNetV2S          0.610           0.610        0.610   
16   EfficientNetV2M          0.384           0.384        0.384   
17   EfficientNetV2L          0.561           0.561        0.561   
18      ConvNeXtTiny          0.776           0.776        0.776   
19     ConvNeXtSmall          0.786           0.786        0.786   
20      ConvNeXtBase          0.815           0.815        0.815   
21     ConvNeXtLarge          0.855           0.855        0.855   
22    ConvNeXtXLarge          0.863           0.863        0.863   

   f1_score train Accuracy test Precision test Recall test f1_score test  
0           0.621         0.595          0.595       0.595         0.595  
1           0.602         0.585          0.585       0.585         0.585  
2           0.622         0.594          0.594       0.594         0.594  
3           0.412         0.397          0.397       0.397         0.397  
4           0.614         0.578          0.578       0.578         0.578  
5           0.386         0.382          0.382       0.382         0.382  
6           0.622         0.590          0.590       0.590         0.590  
7           0.379         0.367          0.367       0.367         0.367  
8           0.205         0.201          0.201       0.201         0.201  
9           0.234         0.230          0.230       0.230         0.230  
10          0.578         0.553          0.553       0.553         0.553  
11          0.557         0.539          0.539       0.539         0.539  
12          0.652         0.631          0.631       0.631         0.631  
13          0.641         0.630          0.630       0.630         0.630  
14          0.593         0.583          0.583       0.583         0.583  
15          0.610         0.601          0.601       0.601         0.601  
16          0.384         0.381          0.381       0.381         0.381  
17          0.561         0.547          0.547       0.547         0.547  
18          0.776         0.740          0.740       0.740         0.740  
19          0.786         0.759          0.759       0.759         0.759  
20          0.815         0.785          0.785       0.785         0.785  
21          0.855         0.829          0.829       0.829         0.829  
22          0.863         0.831          0.831       0.831         0.831  

train_specific_model() Method

• You can train your dataset on a specific model using this method.
• You just have to specify the model_name.
• The method will return the evaluation metrics of the trained model.

Arguments:

• x_train : x_train lists the images for train datasets without labels.
• y_train : y_train holds the labels for train datasets.
• x_test : x_test has the images for test dataset without labels.
• y_test : y_test has the labels for the test dataset.
• model_name
• num_classes : Takes the number of classes for classification. Default value is set to 2.
• batch_size : Trains the image dataset into batches. Defaults Batch size value is 32.
• epochs : Takes the number of number of epochs for the training. Default value set to 1.
• learning_rate : Sets the learning rate for training the model. Default value set to 1e-4.
• momentum : Sets the value of momentum for training the model. Default value set to 0.9.
• SAVE_MODEL : This is a boolean argument which tells the method to save the model and model weight to file or not. The default value is set to False.

Return:

• pandas.dataframe that holds the evaluation metrics of the specific model on the image dataset.

Example:

df_VGG16 = model.train_specific_model( x_train, y_train, x_test, y_test, model_name, num_classes=10, batch_size=32, epochs=20, learning_rate=1e-4, momentum=0.9, SAVE_MODEL = True)

Output:

1563/1563 [==============================] - 37s 17ms/step - loss: 0.3047 - accuracy: 0.3329
Saved model to disk
1563/1563 [==============================] - 12s 7ms/step
313/313 [==============================] - 2s 7ms/step

Evaluation

• You can easily load the specific model and the weights from the disk and evaluate using test dataset.
• You can evaluate the model using two simple methods as:
  • evaluate_saved_model() Method
  • predict_image_saved_model() Method.

evaluate_saved_model() Method

• This method simple tells you the accuracy of the saved model.

Arguments:

• x_test : x_test has the images for test dataset without labels.
• y_test : y_test has the labels for the test dataset.

Example:

model.evaluate_saved_model(x_test, y_test)

Output:

Loaded model from disk
accuracy: 53.04%

predict_image_saved_model() Method

Arguments

• image_path : You have to provide the image path of your single image that you want to evaluate.
• image_width : You have to provide the image width on which your specific model is trained on.
• image_height : You have to provide the image height on which your specific model is trained on.

Example:

image_path  = '/content/maple.jpg'
image_width = 32
image_height = 32
prediction = model.predict_image_saved_model(image_path, image_width, image_height)

• In this example we are providing the image_width and image_height as both 32, as our model is trained on 32x32 images.

Output:

Loaded model from disk

1/1 [==============================] - 0s 334ms/step
[[0.09631938 0.08813578 0.02288193 0.03976142 0.07516313 0.1624043
  0.04893952 0.16084285 0.04286739 0.26268432]]

• You can use this prediction for further evaluation like you can use argmax() method to get the highest value and also you can map the value with label names.

Credits

This software uses the following open source packages:

• Tensorflow
• Keras
• Scikit-learn
• Numpy
• Pandas
• Emojis are taken from here
• highlight.js

Related

LazyPredict - Classification using basic models

Support

Or

You may also like...

• Signature Fraud Detection - A Django app to detect signature fraud from Bank checks.
• AIST - Artificial Intelligence Surveillance Tower
• Parliament Bhaban - A fun OpenGL project for 2D animation.
• Mancala - A simple AI Mancala game built in Python

License

MIT

---

rafsunsheikh.github.io/Web  ·  GitHub @rafsunsheikh  ·  Twitter @RafsunSheikh  ·  LinkedIn @md-rafsun-sheikh