06 - Transfer Learning and Fine-tuning

Transfer learning leverages knowledge from models pre-trained on large datasets to solve new tasks with less data and training time. Instead of training a CNN from scratch, we reuse learned features from models trained on millions of images, then adapt them to our specific problem. This approach often achieves better results with significantly less computational cost.

We will use pre-trained models from Keras Applications to classify flower images, demonstrating both feature extraction and fine-tuning strategies.

Table of Contents

1. Setup and Data Preparation

2. What is Transfer Learning?

3. Pre-trained Models in Keras

4. Feature Extraction with Frozen Backbone

5. Combining Custom CNN with Pre-trained Features

6. Fine-tuning

7. Comparing Approaches

8. Recap

Setup and Data Preparation

First, let’s set up our environment and load the same flowers dataset we used previously.

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
import numpy as np
import matplotlib.pyplot as plt
import pathlib
import os

# Set random seed for reproducibility
tf.random.set_seed(42)
np.random.seed(42)

%matplotlib inline

Loading the Flowers Dataset

# Download dataset (~218MB)
dataset_url = 'https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz'
data_dir = tf.keras.utils.get_file('flower_photos', origin=dataset_url, untar=True)
data_dir = pathlib.Path(data_dir)

# Handle potential nested directory structure
contents = os.listdir(data_dir)
if 'flower_photos' in contents and len(contents) == 1:
    data_dir = os.path.join(data_dir, 'flower_photos')
    data_dir = pathlib.Path(data_dir)

print(f'Total images: {len(list(data_dir.glob("*/*.jpg")))}')

Total images: 3670

Creating Datasets

We use the same 96×96 image size and 80/20 train/validation split.

image_size = (96, 96)
batch_size = 64

# Create training dataset (80% of data)
train_ds = tf.keras.preprocessing.image_dataset_from_directory(
    data_dir,
    validation_split=0.2,
    subset='training',
    seed=42,
    image_size=image_size,
    batch_size=batch_size
)

# Create validation dataset (20% of data)
val_ds = tf.keras.preprocessing.image_dataset_from_directory(
    data_dir,
    validation_split=0.2,
    subset='validation',
    seed=42,
    image_size=image_size,
    batch_size=batch_size
)

class_names = train_ds.class_names
num_classes = len(class_names)
print(f'Classes: {class_names}')

Found 7340 files belonging to 6 classes.
Using 5872 files for training.
Found 7340 files belonging to 6 classes.
Using 1468 files for validation.
Classes: ['daisy', 'dandelion', 'flower_photos', 'roses', 'sunflowers', 'tulips']

Dataset Performance Optimization

Configure datasets for efficient training with caching and prefetching.

# Optimize dataset performance - prevents GPU from waiting for data
AUTOTUNE = tf.data.AUTOTUNE

train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)

print('Dataset optimization applied')

Dataset optimization applied

What is Transfer Learning?

Transfer learning is the practice of using knowledge learned from one task to improve performance on a related task.

The Core Idea

Instead of training a CNN from scratch:

1. Start with a pre-trained model: Trained on a large dataset (e.g., ImageNet with 1.4M images)

2. Reuse learned features: Early layers detect edges, textures, shapes—useful for many vision tasks

3. Adapt to new task: Replace or retrain only the final classification layers

Key Benefits

• Less training data needed: Pre-trained features work well even with small datasets

• Faster training: Only train a small portion of the network

• Better performance: Often achieves higher accuracy than training from scratch

• Lower computational cost: No need for weeks of training on expensive GPUs

When Does Transfer Learning Work?

Transfer learning is most effective when:

• Similar domains: Source and target tasks both involve images

• Related features: Low-level features (edges, textures) transfer well

• Limited target data: Small datasets benefit most from pre-trained knowledge

Two Main Strategies

1. Feature extraction: Freeze the pre-trained backbone, train only new layers

2. Fine-tuning: Unfreeze some layers and train with a small learning rate

Pre-trained Models in Keras

Keras provides access to state-of-the-art CNN architectures pre-trained on ImageNet through tf.keras.applications.

Available Models

Some popular options include:

Model	Parameters	Speed	Accuracy	Best For
MobileNetV2	3.5M	Fast	Good	Mobile/embedded devices
EfficientNetB0	5.3M	Fast	Better	Balanced accuracy/speed
ResNet50	25.6M	Medium	High	When accuracy matters
VGG16	138M	Slow	Good	Simple architecture
InceptionV3	23.9M	Medium	High	Multi-scale features

MobileNetV2

We’ll use MobileNetV2 as our backbone:

• Designed for mobile and embedded devices

• Excellent balance of accuracy and speed

• Relatively few parameters (faster to fine-tune)

• Pre-trained on ImageNet (1000 classes of everyday objects)

ImageNet Pre-training

ImageNet is a large visual database with:

• 1.4 million training images

• 1000 classes (animals, plants, vehicles, objects)

• Many natural images and outdoor scenes

The learned features from ImageNet transfer well to many vision tasks, including flower classification.

Feature Extraction with Frozen Backbone

The first transfer learning strategy is feature extraction: use the pre-trained model as a fixed feature extractor.

Loading the Pre-trained Model

We load MobileNetV2 with ImageNet weights, excluding the top classification layer.

# Load MobileNetV2 without the classification head (include_top=False removes final Dense layer)
base_model = tf.keras.applications.MobileNetV2(
    input_shape=image_size + (3,),
    include_top=False,  # Exclude final classification layer
    weights='imagenet'  # Use ImageNet pre-trained weights
)

print(f'Base model layers: {len(base_model.layers)}')
print(f'Base model output shape: {base_model.output_shape}')

Base model layers: 154
Base model output shape: (None, 3, 3, 1280)

Freezing the Backbone

We freeze the base model to prevent its weights from being updated during training. This preserves the learned ImageNet features.

# Freeze the base model - prevents weights from updating during training
base_model.trainable = False

print(f'Trainable weights after freezing: {len(base_model.trainable_weights)}')
print(f'Non-trainable weights after freezing: {len(base_model.non_trainable_weights)}')

Trainable weights after freezing: 0
Non-trainable weights after freezing: 260

Understanding Model-Specific Preprocessing

Critical: Each pre-trained model expects inputs preprocessed in a specific way. Using incorrect preprocessing will result in poor performance because the model was trained with specific input ranges and transformations.

Different models use different preprocessing strategies:

• MobileNetV2: Scales pixel values from [0, 255] to [-1, 1] range

• ResNet/VGG: Subtracts ImageNet mean [103.939, 116.779, 123.68] from RGB channels

• EfficientNet: Scales to [0, 1] then normalizes with ImageNet statistics

• InceptionV3: Scales to [-1, 1] range

Always use the model’s preprocessing function (tf.keras.applications.{model_name}.preprocess_input) to ensure correct input format.

# Get MobileNetV2-specific preprocessing function
preprocess_input = tf.keras.applications.mobilenet_v2.preprocess_input

# This function scales pixel values to [-1, 1] range (MobileNetV2's expected input)

Building the Transfer Learning Model

We build a new model that:

1. Takes images as input

2. Preprocesses them for MobileNetV2

3. Passes through the frozen backbone (feature extraction)

4. Pools features to a single vector per channel

5. Classifies using new dense layers

# Build the transfer learning model using Functional API
inputs = keras.Input(shape=image_size + (3,))

# Preprocessing: scale to [-1, 1] range expected by MobileNetV2
x = preprocess_input(inputs)

# Pre-trained backbone (frozen): extract high-level features
# training=False keeps batch normalization layers in inference mode
x = base_model(x, training=False)

# Classifier head: convert features to class predictions
x = layers.GlobalAveragePooling2D()(x)  # More efficient than Flatten, reduces parameters
x = layers.Dropout(0.5)(x)  # Regularization to prevent overfitting
outputs = layers.Dense(num_classes, activation='softmax')(x)  # Final classification

model_transfer = keras.Model(inputs, outputs, name='transfer_learning_model')

Model Summary

model_transfer.summary()

# Count trainable vs non-trainable parameters
trainable_count = sum([tf.size(w).numpy() for w in model_transfer.trainable_weights])
non_trainable_count = sum([tf.size(w).numpy() for w in model_transfer.non_trainable_weights])

print(f'\nTrainable parameters: {trainable_count:,}')
print(f'Non-trainable parameters: {non_trainable_count:,}')
print(f'Total parameters: {trainable_count + non_trainable_count:,}')

Model: "transfer_learning_model"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ input_layer_1 (InputLayer)      │ (None, 96, 96, 3)      │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ true_divide (TrueDivide)        │ (None, 96, 96, 3)      │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ subtract (Subtract)             │ (None, 96, 96, 3)      │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ mobilenetv2_1.00_96             │ (None, 3, 3, 1280)     │     2,257,984 │
│ (Functional)                    │                        │               │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ global_average_pooling2d        │ (None, 1280)           │             0 │
│ (GlobalAveragePooling2D)        │                        │               │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout (Dropout)               │ (None, 1280)           │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense (Dense)                   │ (None, 6)              │         7,686 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 2,265,670 (8.64 MB)

 Trainable params: 7,686 (30.02 KB)

 Non-trainable params: 2,257,984 (8.61 MB)

Trainable parameters: 7,686
Non-trainable parameters: 2,257,984
Total parameters: 2,265,670

Observations:

• Most parameters are frozen (non-trainable) - preserves ImageNet knowledge

• Only the new classification layers are trainable - adapts to flower classes

• Training will be fast since we update very few parameters

Training the Model

We compile and train with the frozen backbone.

model_transfer.compile(
    optimizer='adam',  # Default learning rate (1e-3) is fine for new layers
    loss=tf.keras.losses.SparseCategoricalCrossentropy(),
    metrics=['accuracy']
)

epochs_initial = 10
history_transfer = model_transfer.fit(
    train_ds,
    validation_data=val_ds,
    epochs=epochs_initial
)

Epoch 1/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 3:24 2s/step - accuracy: 0.1875 - loss: 2.6627
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.2127 - loss: 2.5061
 4/92 ━━━━━━━━━━━━━━━━━━━━ 4s 49ms/step - accuracy: 0.2230 - loss: 2.4810
 6/92 ━━━━━━━━━━━━━━━━━━━━ 4s 49ms/step - accuracy: 0.2352 - loss: 2.4618
 8/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.2473 - loss: 2.4305
10/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.2598 - loss: 2.3935
12/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.2708 - loss: 2.3624
14/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.2802 - loss: 2.3352
16/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.2878 - loss: 2.3133
18/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.2946 - loss: 2.2921
20/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.3000 - loss: 2.2732
22/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.3051 - loss: 2.2547
24/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.3093 - loss: 2.2364
26/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.3134 - loss: 2.2187
28/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.3174 - loss: 2.2015
30/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.3209 - loss: 2.1856
31/92 ━━━━━━━━━━━━━━━━━━━━ 3s 50ms/step - accuracy: 0.3225 - loss: 2.1782
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3253 - loss: 2.1639
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3278 - loss: 2.1506
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3302 - loss: 2.1379
39/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3324 - loss: 2.1257
40/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3335 - loss: 2.1197
42/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3356 - loss: 2.1081
44/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3374 - loss: 2.0973
46/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3392 - loss: 2.0870
48/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3411 - loss: 2.0769
50/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.3428 - loss: 2.0673
52/92 ━━━━━━━━━━━━━━━━━━━━ 1s 49ms/step - accuracy: 0.3444 - loss: 2.0582
54/92 ━━━━━━━━━━━━━━━━━━━━ 1s 49ms/step - accuracy: 0.3459 - loss: 2.0492
56/92 ━━━━━━━━━━━━━━━━━━━━ 1s 49ms/step - accuracy: 0.3473 - loss: 2.0406
58/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3487 - loss: 2.0318
60/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3502 - loss: 2.0231
61/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3510 - loss: 2.0188
63/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3524 - loss: 2.0105
65/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3539 - loss: 2.0022
67/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3554 - loss: 1.9942
69/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3569 - loss: 1.9862
71/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.3583 - loss: 1.9787
73/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3596 - loss: 1.9714
75/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3609 - loss: 1.9643
76/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3616 - loss: 1.9607
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3629 - loss: 1.9537
80/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3641 - loss: 1.9469
82/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3653 - loss: 1.9403
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3664 - loss: 1.9339
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3675 - loss: 1.9277
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3686 - loss: 1.9217
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3696 - loss: 1.9157
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.3706 - loss: 1.9098
92/92 ━━━━━━━━━━━━━━━━━━━━ 8s 67ms/step - accuracy: 0.4159 - loss: 1.6438 - val_accuracy: 0.4734 - val_loss: 1.1655
Epoch 2/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.5000 - loss: 1.1795
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.4757 - loss: 1.2425
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.4678 - loss: 1.2674
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4583 - loss: 1.2897
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4575 - loss: 1.2974
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4584 - loss: 1.3054
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4584 - loss: 1.3138
15/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4584 - loss: 1.3175
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4580 - loss: 1.3218
18/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4575 - loss: 1.3244
20/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4564 - loss: 1.3278
22/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4552 - loss: 1.3313
24/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4543 - loss: 1.3342
26/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4531 - loss: 1.3367
28/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4522 - loss: 1.3380
30/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4516 - loss: 1.3386
32/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4512 - loss: 1.3389
34/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4510 - loss: 1.3388
36/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4508 - loss: 1.3384
38/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4507 - loss: 1.3381
40/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4505 - loss: 1.3381
42/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4503 - loss: 1.3382
44/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4502 - loss: 1.3383
46/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4500 - loss: 1.3386
48/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4498 - loss: 1.3386
50/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4496 - loss: 1.3386
52/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4495 - loss: 1.3383
54/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4494 - loss: 1.3379
55/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4494 - loss: 1.3376
56/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4495 - loss: 1.3373
57/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4495 - loss: 1.3370
59/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4496 - loss: 1.3362
60/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4496 - loss: 1.3357
62/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4498 - loss: 1.3347
64/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4499 - loss: 1.3336
66/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4500 - loss: 1.3327
68/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4502 - loss: 1.3318
70/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4504 - loss: 1.3306
72/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4505 - loss: 1.3297
74/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4507 - loss: 1.3287
76/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4508 - loss: 1.3279
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4509 - loss: 1.3271
80/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4510 - loss: 1.3264
82/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4511 - loss: 1.3258
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4512 - loss: 1.3253
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4512 - loss: 1.3248
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4513 - loss: 1.3242
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4514 - loss: 1.3236
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4515 - loss: 1.3230
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4566 - loss: 1.2954 - val_accuracy: 0.4278 - val_loss: 1.1088
Epoch 3/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 49ms/step - accuracy: 0.5000 - loss: 1.1471
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.4861 - loss: 1.2000
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.4895 - loss: 1.1914
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4901 - loss: 1.1961
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4885 - loss: 1.1950
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4850 - loss: 1.1970
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4836 - loss: 1.1956
14/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4828 - loss: 1.1954
16/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4812 - loss: 1.1947
18/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4804 - loss: 1.1945
20/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4800 - loss: 1.1941
22/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4802 - loss: 1.1932
24/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4806 - loss: 1.1920
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4809 - loss: 1.1915
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4815 - loss: 1.1902
29/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4819 - loss: 1.1891
30/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4821 - loss: 1.1888
32/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4823 - loss: 1.1886
34/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4828 - loss: 1.1879
36/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4832 - loss: 1.1870
38/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4832 - loss: 1.1868
40/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4831 - loss: 1.1866
42/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4828 - loss: 1.1867
44/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4825 - loss: 1.1867
46/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4823 - loss: 1.1867
48/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4819 - loss: 1.1870
50/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4817 - loss: 1.1871
52/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4814 - loss: 1.1872
54/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4812 - loss: 1.1870
56/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4810 - loss: 1.1871
58/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4807 - loss: 1.1872
60/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4804 - loss: 1.1872
62/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4802 - loss: 1.1870
64/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4800 - loss: 1.1869
66/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4797 - loss: 1.1868
68/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4795 - loss: 1.1868
70/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4792 - loss: 1.1869
72/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4789 - loss: 1.1870
73/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4788 - loss: 1.1870
75/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4785 - loss: 1.1870
77/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4782 - loss: 1.1871
79/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4779 - loss: 1.1871
81/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4776 - loss: 1.1872
83/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4773 - loss: 1.1872
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4770 - loss: 1.1873
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4767 - loss: 1.1873
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4764 - loss: 1.1873
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4761 - loss: 1.1875
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4632 - loss: 1.1926 - val_accuracy: 0.4298 - val_loss: 1.1221
Epoch 4/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 51ms/step - accuracy: 0.4531 - loss: 1.0588
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.4271 - loss: 1.1937
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.4261 - loss: 1.1953
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4256 - loss: 1.1951
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4297 - loss: 1.1901
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4317 - loss: 1.1892
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4343 - loss: 1.1870
15/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4376 - loss: 1.1844
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4415 - loss: 1.1802
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4438 - loss: 1.1765
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4462 - loss: 1.1723
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4481 - loss: 1.1687
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4498 - loss: 1.1660
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4510 - loss: 1.1640
29/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4522 - loss: 1.1620
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4534 - loss: 1.1599
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4543 - loss: 1.1579
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4553 - loss: 1.1556
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4565 - loss: 1.1534
38/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4571 - loss: 1.1523
39/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4576 - loss: 1.1514
41/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4586 - loss: 1.1495
43/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4596 - loss: 1.1476
45/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4606 - loss: 1.1459
47/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4615 - loss: 1.1445
49/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4622 - loss: 1.1433
51/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4629 - loss: 1.1421
53/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4634 - loss: 1.1411
55/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4639 - loss: 1.1401
57/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4643 - loss: 1.1392
59/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4648 - loss: 1.1381
61/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4653 - loss: 1.1370
63/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4658 - loss: 1.1359
65/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4661 - loss: 1.1350
67/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4664 - loss: 1.1343
69/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4667 - loss: 1.1336
71/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4670 - loss: 1.1329
73/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4672 - loss: 1.1323
75/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4674 - loss: 1.1319
77/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4675 - loss: 1.1314
79/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4676 - loss: 1.1311
81/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4677 - loss: 1.1307
83/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4677 - loss: 1.1304
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4678 - loss: 1.1302
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4678 - loss: 1.1299
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4679 - loss: 1.1297
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4679 - loss: 1.1296
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4690 - loss: 1.1244 - val_accuracy: 0.4149 - val_loss: 1.1264
Epoch 5/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.4375 - loss: 1.1789
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.4505 - loss: 1.1699
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.4702 - loss: 1.1344
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4762 - loss: 1.1282
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4815 - loss: 1.1230
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4839 - loss: 1.1174
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4853 - loss: 1.1118
15/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4857 - loss: 1.1074
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4862 - loss: 1.1047
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4869 - loss: 1.1028
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4872 - loss: 1.1015
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4878 - loss: 1.1002
24/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4882 - loss: 1.0991
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4884 - loss: 1.0984
26/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4888 - loss: 1.0974
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4889 - loss: 1.0966
29/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4892 - loss: 1.0950
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.4894 - loss: 1.0936
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4896 - loss: 1.0922
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4897 - loss: 1.0910
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4899 - loss: 1.0896
38/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.4899 - loss: 1.0890
40/92 ━━━━━━━━━━━━━━━━━━━━ 2s 49ms/step - accuracy: 0.4899 - loss: 1.0879
42/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4899 - loss: 1.0869
44/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4901 - loss: 1.0859
46/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4902 - loss: 1.0851
47/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4903 - loss: 1.0847
49/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4903 - loss: 1.0841
51/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4904 - loss: 1.0835
53/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4904 - loss: 1.0830
55/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4904 - loss: 1.0825
57/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4904 - loss: 1.0820
59/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4903 - loss: 1.0816
61/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4903 - loss: 1.0812
63/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4903 - loss: 1.0809
65/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4902 - loss: 1.0807
67/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4901 - loss: 1.0807
69/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4899 - loss: 1.0807
71/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4898 - loss: 1.0807
73/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.4896 - loss: 1.0808
75/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.4894 - loss: 1.0808
77/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.4891 - loss: 1.0810
79/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4889 - loss: 1.0811
81/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4887 - loss: 1.0812
83/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4885 - loss: 1.0811
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4884 - loss: 1.0811
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4882 - loss: 1.0810
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4880 - loss: 1.0809
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4878 - loss: 1.0809
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4794 - loss: 1.0784 - val_accuracy: 0.3985 - val_loss: 1.1180
Epoch 6/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.4375 - loss: 1.0949
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.4592 - loss: 1.0643
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.4852 - loss: 1.0224
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4921 - loss: 1.0177
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4921 - loss: 1.0217
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4904 - loss: 1.0279
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4883 - loss: 1.0322
15/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4870 - loss: 1.0329
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4860 - loss: 1.0334
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4862 - loss: 1.0337
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4871 - loss: 1.0332
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.4886 - loss: 1.0315
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4904 - loss: 1.0291
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.4916 - loss: 1.0271
29/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4927 - loss: 1.0255
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4936 - loss: 1.0243
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4943 - loss: 1.0237
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4949 - loss: 1.0232
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4955 - loss: 1.0227
39/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4959 - loss: 1.0225
41/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4961 - loss: 1.0224
43/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4962 - loss: 1.0228
45/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4962 - loss: 1.0230
47/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.4962 - loss: 1.0235
48/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.4961 - loss: 1.0237
50/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0240
52/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0243
54/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0246
56/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0247
58/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0249
60/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0250
62/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0251
64/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4959 - loss: 1.0253
66/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4959 - loss: 1.0255
68/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0257
70/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4960 - loss: 1.0258
72/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4961 - loss: 1.0259
74/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4961 - loss: 1.0260
76/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4962 - loss: 1.0262
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4962 - loss: 1.0263
80/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4962 - loss: 1.0265
82/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4962 - loss: 1.0267
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4961 - loss: 1.0269
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4960 - loss: 1.0272
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4959 - loss: 1.0275
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4958 - loss: 1.0279
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4956 - loss: 1.0283
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4879 - loss: 1.0464 - val_accuracy: 0.3822 - val_loss: 1.1081
Epoch 7/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.4844 - loss: 0.8474
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.5069 - loss: 0.8889
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.5032 - loss: 0.9170
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5008 - loss: 0.9341
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.5020 - loss: 0.9392
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.5018 - loss: 0.9443
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5022 - loss: 0.9470
15/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5030 - loss: 0.9515
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.5040 - loss: 0.9552
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5048 - loss: 0.9590
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5053 - loss: 0.9615
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5056 - loss: 0.9633
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5058 - loss: 0.9650
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5059 - loss: 0.9662
29/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5060 - loss: 0.9676
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5058 - loss: 0.9693
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5057 - loss: 0.9708
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5053 - loss: 0.9722
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5050 - loss: 0.9735
39/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5049 - loss: 0.9743
41/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5047 - loss: 0.9752
43/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5043 - loss: 0.9763
45/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5038 - loss: 0.9775
47/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5035 - loss: 0.9786
49/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5031 - loss: 0.9796
51/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5027 - loss: 0.9805
53/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5023 - loss: 0.9813
55/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5019 - loss: 0.9822
57/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5015 - loss: 0.9830
59/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5011 - loss: 0.9838
61/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5007 - loss: 0.9846
62/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5005 - loss: 0.9850
64/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5002 - loss: 0.9856
66/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5000 - loss: 0.9863
68/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4997 - loss: 0.9870
70/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.4994 - loss: 0.9877
72/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4991 - loss: 0.9883
74/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4989 - loss: 0.9888
76/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4987 - loss: 0.9893
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4985 - loss: 0.9899
80/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4983 - loss: 0.9905
82/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4981 - loss: 0.9910
83/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4980 - loss: 0.9913
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4978 - loss: 0.9918
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4976 - loss: 0.9924
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4974 - loss: 0.9929
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.4972 - loss: 0.9934
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4898 - loss: 1.0131 - val_accuracy: 0.3787 - val_loss: 1.1095
Epoch 8/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.4375 - loss: 1.0026
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.4714 - loss: 1.0030
 4/92 ━━━━━━━━━━━━━━━━━━━━ 4s 49ms/step - accuracy: 0.4844 - loss: 0.9951
 6/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.4886 - loss: 1.0015
 8/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.4914 - loss: 1.0070
10/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4925 - loss: 1.0106
12/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4930 - loss: 1.0117
14/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4930 - loss: 1.0132
16/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4929 - loss: 1.0128
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4930 - loss: 1.0122
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4943 - loss: 1.0095
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4961 - loss: 1.0067
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4975 - loss: 1.0042
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 49ms/step - accuracy: 0.4983 - loss: 1.0023
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4985 - loss: 1.0015
29/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.4989 - loss: 1.0004
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4988 - loss: 0.9999
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4985 - loss: 0.9997
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4979 - loss: 0.9996
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4974 - loss: 0.9993
39/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4970 - loss: 0.9990
41/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4965 - loss: 0.9987
43/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4962 - loss: 0.9983
45/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4960 - loss: 0.9979
47/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4959 - loss: 0.9975
49/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.4958 - loss: 0.9971
51/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4956 - loss: 0.9970
52/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4955 - loss: 0.9969
54/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4954 - loss: 0.9968
56/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4953 - loss: 0.9968
58/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4952 - loss: 0.9968
60/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4951 - loss: 0.9968
62/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4951 - loss: 0.9968
63/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4951 - loss: 0.9967
65/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4952 - loss: 0.9966
67/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4952 - loss: 0.9965
69/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4952 - loss: 0.9965
71/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.4952 - loss: 0.9965
73/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.4953 - loss: 0.9966
75/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.4952 - loss: 0.9967
77/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4952 - loss: 0.9968
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4952 - loss: 0.9969
79/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4953 - loss: 0.9969
80/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4953 - loss: 0.9969
81/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4953 - loss: 0.9970
82/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4953 - loss: 0.9971
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4952 - loss: 0.9973
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4952 - loss: 0.9975
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4952 - loss: 0.9977
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4952 - loss: 0.9979
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - accuracy: 0.4951 - loss: 0.9980
92/92 ━━━━━━━━━━━━━━━━━━━━ 6s 61ms/step - accuracy: 0.4946 - loss: 1.0050 - val_accuracy: 0.4080 - val_loss: 1.1515
Epoch 9/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.4688 - loss: 1.2861
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.5217 - loss: 1.1225
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 47ms/step - accuracy: 0.5244 - loss: 1.0871
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5284 - loss: 1.0593
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5324 - loss: 1.0393
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.5353 - loss: 1.0244
12/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5365 - loss: 1.0181
14/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.5363 - loss: 1.0106
16/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5355 - loss: 1.0052
18/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5342 - loss: 1.0007
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.5335 - loss: 0.9989
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 48ms/step - accuracy: 0.5319 - loss: 0.9954
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5302 - loss: 0.9931
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5283 - loss: 0.9919
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5267 - loss: 0.9909
29/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5253 - loss: 0.9901
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5241 - loss: 0.9894
32/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5234 - loss: 0.9892
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5229 - loss: 0.9889
34/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5224 - loss: 0.9884
36/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5216 - loss: 0.9874
38/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5210 - loss: 0.9864
40/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5205 - loss: 0.9855
42/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5199 - loss: 0.9851
44/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5195 - loss: 0.9846
46/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5190 - loss: 0.9845
48/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5186 - loss: 0.9843
50/92 ━━━━━━━━━━━━━━━━━━━━ 2s 48ms/step - accuracy: 0.5183 - loss: 0.9842
52/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5179 - loss: 0.9840
54/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5175 - loss: 0.9838
55/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5173 - loss: 0.9837
56/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5172 - loss: 0.9836
58/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5168 - loss: 0.9835
60/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5165 - loss: 0.9834
61/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5163 - loss: 0.9835
62/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5161 - loss: 0.9835
64/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5157 - loss: 0.9836
66/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5154 - loss: 0.9837
68/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5150 - loss: 0.9838
70/92 ━━━━━━━━━━━━━━━━━━━━ 1s 48ms/step - accuracy: 0.5146 - loss: 0.9839
72/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5141 - loss: 0.9840
74/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5137 - loss: 0.9842
76/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5133 - loss: 0.9844
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5129 - loss: 0.9846
80/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5124 - loss: 0.9849
81/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5122 - loss: 0.9851
82/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5120 - loss: 0.9853
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5115 - loss: 0.9855
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5113 - loss: 0.9857
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5109 - loss: 0.9860
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5104 - loss: 0.9863
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - accuracy: 0.5100 - loss: 0.9866
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 60ms/step - accuracy: 0.4908 - loss: 1.0006 - val_accuracy: 0.3488 - val_loss: 1.1116
Epoch 10/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 4s 48ms/step - accuracy: 0.5312 - loss: 1.0082
 3/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.5347 - loss: 0.9522
 5/92 ━━━━━━━━━━━━━━━━━━━━ 4s 46ms/step - accuracy: 0.5321 - loss: 0.9432
 7/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.5306 - loss: 0.9384
 9/92 ━━━━━━━━━━━━━━━━━━━━ 3s 46ms/step - accuracy: 0.5278 - loss: 0.9379
11/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5259 - loss: 0.9383
13/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5245 - loss: 0.9395
15/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5234 - loss: 0.9405
17/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5221 - loss: 0.9429
19/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5215 - loss: 0.9443
21/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5210 - loss: 0.9462
23/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5201 - loss: 0.9493
25/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5194 - loss: 0.9516
27/92 ━━━━━━━━━━━━━━━━━━━━ 3s 47ms/step - accuracy: 0.5187 - loss: 0.9539
29/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5178 - loss: 0.9562
31/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5172 - loss: 0.9579
33/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5166 - loss: 0.9594
35/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5160 - loss: 0.9608
37/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5157 - loss: 0.9620
39/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5154 - loss: 0.9630
41/92 ━━━━━━━━━━━━━━━━━━━━ 2s 47ms/step - accuracy: 0.5153 - loss: 0.9639
43/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5151 - loss: 0.9648
45/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5148 - loss: 0.9656
47/92 ━━━━━━━━━━━━━━━━━━━━ 2s 46ms/step - accuracy: 0.5144 - loss: 0.9665
49/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5140 - loss: 0.9673
51/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5136 - loss: 0.9680
53/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5131 - loss: 0.9687
55/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5127 - loss: 0.9694
57/92 ━━━━━━━━━━━━━━━━━━━━ 1s 46ms/step - accuracy: 0.5122 - loss: 0.9701
59/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5118 - loss: 0.9707
61/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5113 - loss: 0.9714
63/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5108 - loss: 0.9720
65/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5103 - loss: 0.9727
67/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5098 - loss: 0.9732
68/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5096 - loss: 0.9734
70/92 ━━━━━━━━━━━━━━━━━━━━ 1s 47ms/step - accuracy: 0.5090 - loss: 0.9740
72/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5086 - loss: 0.9744
74/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5083 - loss: 0.9747
76/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5080 - loss: 0.9751
78/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5077 - loss: 0.9754
79/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5076 - loss: 0.9756
81/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5073 - loss: 0.9760
83/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5070 - loss: 0.9763
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5068 - loss: 0.9767
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5065 - loss: 0.9771
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5064 - loss: 0.9773
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5061 - loss: 0.9777
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - accuracy: 0.5058 - loss: 0.9781
92/92 ━━━━━━━━━━━━━━━━━━━━ 5s 59ms/step - accuracy: 0.4942 - loss: 0.9979 - val_accuracy: 0.3529 - val_loss: 1.1607

Visualizing Results

import pandas as pd

def plot_training_history(history, title='Training History'):
    """Plot training and validation metrics."""
    hist = pd.DataFrame(history.history)
    hist['epoch'] = history.epoch

    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))

    # Plot loss
    ax1.plot(hist['epoch'], hist['loss'], label='Train Loss', marker='o')
    ax1.plot(hist['epoch'], hist['val_loss'], label='Val Loss', marker='s')
    ax1.set_xlabel('Epoch')
    ax1.set_ylabel('Loss')
    ax1.set_title(f'{title} - Loss')
    ax1.legend()
    ax1.grid(True, alpha=0.3)

    # Plot accuracy
    ax2.plot(hist['epoch'], hist['accuracy'], label='Train Accuracy', marker='o')
    ax2.plot(hist['epoch'], hist['val_accuracy'], label='Val Accuracy', marker='s')
    ax2.set_xlabel('Epoch')
    ax2.set_ylabel('Accuracy')
    ax2.set_title(f'{title} - Accuracy')
    ax2.legend()
    ax2.grid(True, alpha=0.3)

    plt.tight_layout()
    plt.show()

plot_training_history(history_transfer, 'Transfer Learning (Frozen Backbone)')

Evaluating Performance

val_loss_transfer, val_acc_transfer = model_transfer.evaluate(val_ds, verbose=0)
print(f'Transfer Learning Results (Frozen Backbone):')
print(f'  Validation Loss: {val_loss_transfer:.4f}')
print(f'  Validation Accuracy: {val_acc_transfer:.4f}')

Transfer Learning Results (Frozen Backbone):
  Validation Loss: 1.1607
  Validation Accuracy: 0.3529

Making Predictions

Let’s visualize some predictions on validation images.

# Get predictions on one batch
for images, labels in val_ds.take(1):
    predictions = model_transfer.predict(images, verbose=0)
    predicted_classes = predictions.argmax(axis=-1)
    
    # Display some examples
    plt.figure(figsize=(15, 6))
    for i in range(6):
        ax = plt.subplot(2, 3, i + 1)
        plt.imshow(images[i].numpy().astype("uint8"))
        
        true_class = class_names[labels[i]]
        pred_class = class_names[predicted_classes[i]]
        confidence = predictions[i][predicted_classes[i]]
        
        # Color code: green if correct, red if wrong
        color = 'green' if true_class == pred_class else 'red'
        plt.title(f'True: {true_class}\nPred: {pred_class} ({confidence:.2f})', color=color)
        plt.axis('off')
    
    plt.tight_layout()
    plt.show()
    break

Combining Custom CNN with Pre-trained Features

A hybrid approach uses both custom convolutional layers and pre-trained features. This is useful when:

• Your dataset has unique characteristics not well represented in ImageNet

• You want to learn task-specific low-level features while leveraging pre-trained high-level features

• You have sufficient data to train custom layers without overfitting

Building a Hybrid Model

We create two parallel branches:

1. Custom CNN branch: Learns task-specific features from scratch

2. Pre-trained branch: Extracts general features from ImageNet

Then concatenate and combine both feature sets for classification.

# Build hybrid model with two parallel branches
inputs = keras.Input(shape=image_size + (3,))
x = preprocess_input(inputs)  # Preprocess for MobileNetV2

# Branch 1: Custom CNN path - learns flower-specific features
custom_branch = layers.Conv2D(32, 3, activation='relu', padding='same')(x)
custom_branch = layers.MaxPooling2D(2)(custom_branch)
custom_branch = layers.Conv2D(64, 3, activation='relu', padding='same')(custom_branch)
custom_branch = layers.GlobalAveragePooling2D()(custom_branch)  # Reduces to 64 features

# Branch 2: Pre-trained features - extracts general visual features
pretrained_branch = base_model(x, training=False)  # Keep frozen
pretrained_branch = layers.GlobalAveragePooling2D()(pretrained_branch)  # Reduces to 1280 features

# Combine both branches: concatenate custom (64) + pretrained (1280) = 1344 features
combined = layers.Concatenate()([custom_branch, pretrained_branch])
combined = layers.Dense(128, activation='relu')(combined)  # Learn feature combinations
combined = layers.Dropout(0.5)(combined)  # Regularization
outputs = layers.Dense(num_classes, activation='softmax')(combined)

model_hybrid = keras.Model(inputs, outputs, name='hybrid_model')

Training the Hybrid Model

model_hybrid.compile(
    optimizer='adam',
    loss=tf.keras.losses.SparseCategoricalCrossentropy(),
    metrics=['accuracy']
)

# Train hybrid model
history_hybrid = model_hybrid.fit(
    train_ds,
    validation_data=val_ds,
    epochs=10
)

# Evaluate
val_loss_hybrid, val_acc_hybrid = model_hybrid.evaluate(val_ds, verbose=0)
print(f'\nHybrid Model Results:')
print(f'  Validation Loss: {val_loss_hybrid:.4f}')
print(f'  Validation Accuracy: {val_acc_hybrid:.4f}')

Epoch 1/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 2:39 2s/step - accuracy: 0.1875 - loss: 2.5249
 2/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.2305 - loss: 2.3740
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.2595 - loss: 2.3162
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.2826 - loss: 2.2675
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.2973 - loss: 2.2334
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.3076 - loss: 2.2009
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.3163 - loss: 2.1694
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.3241 - loss: 2.1401
 9/92 ━━━━━━━━━━━━━━━━━━━━ 10s 123ms/step - accuracy: 0.3298 - loss: 2.1145
10/92 ━━━━━━━━━━━━━━━━━━━━ 10s 123ms/step - accuracy: 0.3338 - loss: 2.0923
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.3371 - loss: 2.0721
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.3397 - loss: 2.0530
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.3425 - loss: 2.0331
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.3452 - loss: 2.0140
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.3474 - loss: 1.9983
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.3499 - loss: 1.9830
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.3521 - loss: 1.9687
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3541 - loss: 1.9554
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3561 - loss: 1.9428
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3581 - loss: 1.9305
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3601 - loss: 1.9186
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3621 - loss: 1.9070
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3643 - loss: 1.8955
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3662 - loss: 1.8844
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.3680 - loss: 1.8737
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.3698 - loss: 1.8634
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.3715 - loss: 1.8533
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.3731 - loss: 1.8435
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.3746 - loss: 1.8341
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.3759 - loss: 1.8250
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.3772 - loss: 1.8163
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.3785 - loss: 1.8076
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.3797 - loss: 1.7991
34/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.3809 - loss: 1.7908
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3820 - loss: 1.7827
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3831 - loss: 1.7749
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3842 - loss: 1.7674
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3852 - loss: 1.7601
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3862 - loss: 1.7530
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3873 - loss: 1.7459
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.3884 - loss: 1.7389
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.3895 - loss: 1.7321
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.3905 - loss: 1.7255
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.3915 - loss: 1.7191
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.3924 - loss: 1.7129
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.3933 - loss: 1.7068
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.3943 - loss: 1.7008
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.3952 - loss: 1.6950
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.3962 - loss: 1.6892
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.3970 - loss: 1.6837
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.3978 - loss: 1.6783
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.3986 - loss: 1.6729
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.3995 - loss: 1.6676
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4002 - loss: 1.6624
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4010 - loss: 1.6573
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4018 - loss: 1.6524
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4025 - loss: 1.6476
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4031 - loss: 1.6429
59/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4038 - loss: 1.6383
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4044 - loss: 1.6338
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4051 - loss: 1.6294
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4056 - loss: 1.6251
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4062 - loss: 1.6209
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4068 - loss: 1.6168
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4073 - loss: 1.6128
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4078 - loss: 1.6088
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4083 - loss: 1.6049
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4088 - loss: 1.6010
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4093 - loss: 1.5973
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4098 - loss: 1.5936
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4102 - loss: 1.5900
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4107 - loss: 1.5864
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4111 - loss: 1.5829
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4116 - loss: 1.5795
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4120 - loss: 1.5762
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4124 - loss: 1.5729
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4129 - loss: 1.5696
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4133 - loss: 1.5664
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4137 - loss: 1.5632
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4142 - loss: 1.5601
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4146 - loss: 1.5570
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4150 - loss: 1.5540
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4153 - loss: 1.5510
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4157 - loss: 1.5481
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4160 - loss: 1.5452
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4163 - loss: 1.5424
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4167 - loss: 1.5396
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4170 - loss: 1.5368
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4173 - loss: 1.5341
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4176 - loss: 1.5314
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4179 - loss: 1.5288
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4182 - loss: 1.5262
92/92 ━━━━━━━━━━━━━━━━━━━━ 15s 143ms/step - accuracy: 0.4441 - loss: 1.2893 - val_accuracy: 0.4646 - val_loss: 1.0155
Epoch 2/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 11s 121ms/step - accuracy: 0.3750 - loss: 1.1632
 2/92 ━━━━━━━━━━━━━━━━━━━━ 10s 117ms/step - accuracy: 0.4297 - loss: 1.1176
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 118ms/step - accuracy: 0.4566 - loss: 1.0874
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 118ms/step - accuracy: 0.4704 - loss: 1.0704
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4801 - loss: 1.0630
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4855 - loss: 1.0560
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4886 - loss: 1.0498
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4917 - loss: 1.0434
 9/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4932 - loss: 1.0390
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4950 - loss: 1.0348
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4956 - loss: 1.0323
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4962 - loss: 1.0300
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4975 - loss: 1.0268
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4987 - loss: 1.0246
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4998 - loss: 1.0223
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5009 - loss: 1.0203
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5015 - loss: 1.0190
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5021 - loss: 1.0183
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5029 - loss: 1.0175
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5037 - loss: 1.0170
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5043 - loss: 1.0165
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5047 - loss: 1.0161
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5050 - loss: 1.0157
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.5052 - loss: 1.0153
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.5054 - loss: 1.0151
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 119ms/step - accuracy: 0.5056 - loss: 1.0146
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5057 - loss: 1.0145
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5058 - loss: 1.0143
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5058 - loss: 1.0142
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5058 - loss: 1.0141
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5058 - loss: 1.0140
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5057 - loss: 1.0138
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5055 - loss: 1.0137
34/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5052 - loss: 1.0137
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5050 - loss: 1.0138
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5048 - loss: 1.0138
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5046 - loss: 1.0138
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5043 - loss: 1.0138
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5040 - loss: 1.0139
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5038 - loss: 1.0140
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5036 - loss: 1.0140
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5035 - loss: 1.0140
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5033 - loss: 1.0141
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5031 - loss: 1.0141
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5030 - loss: 1.0141
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5029 - loss: 1.0140
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5028 - loss: 1.0138
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5027 - loss: 1.0137
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5027 - loss: 1.0135
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5026 - loss: 1.0133
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5026 - loss: 1.0132
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5026 - loss: 1.0130
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5026 - loss: 1.0129
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5026 - loss: 1.0127
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5026 - loss: 1.0126
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5026 - loss: 1.0125
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5026 - loss: 1.0125
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5025 - loss: 1.0124
59/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5025 - loss: 1.0124
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5024 - loss: 1.0123
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5023 - loss: 1.0123
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5022 - loss: 1.0123
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5021 - loss: 1.0122
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5019 - loss: 1.0122
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5018 - loss: 1.0122
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5017 - loss: 1.0121
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5015 - loss: 1.0121
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5013 - loss: 1.0121
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5011 - loss: 1.0120
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5009 - loss: 1.0120
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5007 - loss: 1.0120
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5005 - loss: 1.0120
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5003 - loss: 1.0120
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5001 - loss: 1.0120
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.4999 - loss: 1.0120
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.4998 - loss: 1.0120
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.4996 - loss: 1.0119
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.4994 - loss: 1.0119
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.4992 - loss: 1.0119
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 123ms/step - accuracy: 0.4990 - loss: 1.0119
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 123ms/step - accuracy: 0.4988 - loss: 1.0120
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 123ms/step - accuracy: 0.4986 - loss: 1.0120
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 123ms/step - accuracy: 0.4984 - loss: 1.0120
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4982 - loss: 1.0120
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4980 - loss: 1.0120
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4978 - loss: 1.0120
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4977 - loss: 1.0119
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4975 - loss: 1.0119
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4974 - loss: 1.0118
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4973 - loss: 1.0118
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4971 - loss: 1.0118
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.4970 - loss: 1.0117
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 139ms/step - accuracy: 0.4833 - loss: 1.0087 - val_accuracy: 0.4257 - val_loss: 0.9903
Epoch 3/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 11s 121ms/step - accuracy: 0.4219 - loss: 1.0957
 2/92 ━━━━━━━━━━━━━━━━━━━━ 10s 118ms/step - accuracy: 0.4570 - loss: 1.0308
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.4592 - loss: 1.0194
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4635 - loss: 1.0108
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.4702 - loss: 0.9994
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.4734 - loss: 0.9932
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.4779 - loss: 0.9870
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.4823 - loss: 0.9808
 9/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4837 - loss: 0.9780
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4855 - loss: 0.9752
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 117ms/step - accuracy: 0.4862 - loss: 0.9733
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 118ms/step - accuracy: 0.4869 - loss: 0.9723
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 118ms/step - accuracy: 0.4876 - loss: 0.9709
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 118ms/step - accuracy: 0.4884 - loss: 0.9697
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 118ms/step - accuracy: 0.4889 - loss: 0.9683
16/92 ━━━━━━━━━━━━━━━━━━━━ 8s 118ms/step - accuracy: 0.4890 - loss: 0.9671
17/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4893 - loss: 0.9662
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4893 - loss: 0.9653
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4892 - loss: 0.9645
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4891 - loss: 0.9638
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4891 - loss: 0.9629
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4891 - loss: 0.9620
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4893 - loss: 0.9614
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.4895 - loss: 0.9608
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.4896 - loss: 0.9602
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4896 - loss: 0.9599
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4897 - loss: 0.9593
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4899 - loss: 0.9587
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4900 - loss: 0.9581
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4902 - loss: 0.9576
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4905 - loss: 0.9569
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4908 - loss: 0.9565
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.4912 - loss: 0.9559
34/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4915 - loss: 0.9555
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4918 - loss: 0.9550
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4921 - loss: 0.9545
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4923 - loss: 0.9541
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4925 - loss: 0.9538
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4927 - loss: 0.9534
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4930 - loss: 0.9531
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4932 - loss: 0.9528
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.4933 - loss: 0.9526
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4934 - loss: 0.9523
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4935 - loss: 0.9521
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4936 - loss: 0.9518
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4937 - loss: 0.9515
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4939 - loss: 0.9512
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4941 - loss: 0.9510
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4942 - loss: 0.9508
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.4943 - loss: 0.9506
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4945 - loss: 0.9504
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4946 - loss: 0.9502
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4948 - loss: 0.9500
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4949 - loss: 0.9498
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4950 - loss: 0.9496
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4950 - loss: 0.9495
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4951 - loss: 0.9493
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.4952 - loss: 0.9492
59/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4953 - loss: 0.9490
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4953 - loss: 0.9489
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4953 - loss: 0.9488
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4954 - loss: 0.9486
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4954 - loss: 0.9485
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4954 - loss: 0.9483
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4955 - loss: 0.9482
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4955 - loss: 0.9481
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.4955 - loss: 0.9479
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4956 - loss: 0.9478
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4956 - loss: 0.9477
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4956 - loss: 0.9476
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4956 - loss: 0.9475
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4956 - loss: 0.9475
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4955 - loss: 0.9474
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4955 - loss: 0.9474
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4954 - loss: 0.9473
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4954 - loss: 0.9473
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4953 - loss: 0.9472
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4952 - loss: 0.9471
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4952 - loss: 0.9471
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4951 - loss: 0.9470
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4950 - loss: 0.9470
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4950 - loss: 0.9469
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4949 - loss: 0.9469
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4948 - loss: 0.9469
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4947 - loss: 0.9468
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4947 - loss: 0.9468
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4946 - loss: 0.9468
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4945 - loss: 0.9468
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4944 - loss: 0.9468
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4944 - loss: 0.9468
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.4943 - loss: 0.9468
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4943 - loss: 0.9467
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 137ms/step - accuracy: 0.4886 - loss: 0.9453 - val_accuracy: 0.4135 - val_loss: 0.9856
Epoch 4/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.4219 - loss: 0.9605
 2/92 ━━━━━━━━━━━━━━━━━━━━ 11s 123ms/step - accuracy: 0.4688 - loss: 0.9256
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4757 - loss: 0.9282
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4798 - loss: 0.9314
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4814 - loss: 0.9361
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4819 - loss: 0.9396
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4809 - loss: 0.9409
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4816 - loss: 0.9407
 9/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4833 - loss: 0.9396
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.4853 - loss: 0.9378
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.4865 - loss: 0.9361
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.4873 - loss: 0.9347
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.4872 - loss: 0.9340
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4878 - loss: 0.9329
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.4882 - loss: 0.9318
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4886 - loss: 0.9307
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4891 - loss: 0.9292
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4898 - loss: 0.9282
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4904 - loss: 0.9273
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4910 - loss: 0.9266
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4917 - loss: 0.9260
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4924 - loss: 0.9254
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4930 - loss: 0.9250
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4938 - loss: 0.9244
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4947 - loss: 0.9236
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4954 - loss: 0.9230
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4960 - loss: 0.9227
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4964 - loss: 0.9226
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4967 - loss: 0.9225
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4970 - loss: 0.9223
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4974 - loss: 0.9221
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4977 - loss: 0.9218
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4982 - loss: 0.9214
34/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4987 - loss: 0.9210
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4991 - loss: 0.9206
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4995 - loss: 0.9203
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4998 - loss: 0.9199
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5002 - loss: 0.9196
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5005 - loss: 0.9192
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5008 - loss: 0.9189
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5011 - loss: 0.9186
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5015 - loss: 0.9182
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5018 - loss: 0.9178
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5020 - loss: 0.9175
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5023 - loss: 0.9172
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5025 - loss: 0.9170
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5026 - loss: 0.9168
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5027 - loss: 0.9166
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5028 - loss: 0.9164
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5029 - loss: 0.9162
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9161
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9159
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9158
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9156
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9155
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9153
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5029 - loss: 0.9151
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5030 - loss: 0.9150
59/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5030 - loss: 0.9148
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5030 - loss: 0.9147
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5031 - loss: 0.9145
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5031 - loss: 0.9143
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5032 - loss: 0.9142
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5032 - loss: 0.9140
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5033 - loss: 0.9138
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5033 - loss: 0.9137
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5033 - loss: 0.9136
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5034 - loss: 0.9134
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5034 - loss: 0.9133
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5034 - loss: 0.9133
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5035 - loss: 0.9132
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5035 - loss: 0.9131
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5035 - loss: 0.9129
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5035 - loss: 0.9129
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5035 - loss: 0.9128
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5035 - loss: 0.9127
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5035 - loss: 0.9126
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5035 - loss: 0.9125
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5035 - loss: 0.9124
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5034 - loss: 0.9123
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5034 - loss: 0.9123
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5034 - loss: 0.9122
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5034 - loss: 0.9122
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.5034 - loss: 0.9121
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.5034 - loss: 0.9120
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5034 - loss: 0.9120
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5033 - loss: 0.9119
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5033 - loss: 0.9118
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5033 - loss: 0.9117
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5033 - loss: 0.9116
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5032 - loss: 0.9116
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5032 - loss: 0.9115
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 138ms/step - accuracy: 0.5017 - loss: 0.9067 - val_accuracy: 0.3856 - val_loss: 1.0035
Epoch 5/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.5312 - loss: 0.7833
 2/92 ━━━━━━━━━━━━━━━━━━━━ 10s 118ms/step - accuracy: 0.5547 - loss: 0.7760
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.5608 - loss: 0.7825
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.5612 - loss: 0.7861
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5546 - loss: 0.7927
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5494 - loss: 0.8004
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5449 - loss: 0.8054
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5437 - loss: 0.8078
 9/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5413 - loss: 0.8104
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5392 - loss: 0.8132
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5379 - loss: 0.8161
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5370 - loss: 0.8189
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5363 - loss: 0.8217
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5354 - loss: 0.8239
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5351 - loss: 0.8254
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5346 - loss: 0.8274
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5341 - loss: 0.8293
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5339 - loss: 0.8308
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5336 - loss: 0.8324
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5334 - loss: 0.8338
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5332 - loss: 0.8351
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5332 - loss: 0.8361
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5332 - loss: 0.8370
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5330 - loss: 0.8381
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5327 - loss: 0.8391
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5325 - loss: 0.8400
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5323 - loss: 0.8408
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5321 - loss: 0.8414
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5319 - loss: 0.8421
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5316 - loss: 0.8429
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5313 - loss: 0.8436
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5308 - loss: 0.8443
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.5305 - loss: 0.8449
34/92 ━━━━━━━━━━━━━━━━━━━━ 7s 122ms/step - accuracy: 0.5302 - loss: 0.8453
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 122ms/step - accuracy: 0.5300 - loss: 0.8457
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5298 - loss: 0.8461
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5296 - loss: 0.8465
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5294 - loss: 0.8468
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5292 - loss: 0.8472
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5290 - loss: 0.8476
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5288 - loss: 0.8480
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5286 - loss: 0.8483
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5284 - loss: 0.8487
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5282 - loss: 0.8491
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5281 - loss: 0.8495
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5279 - loss: 0.8500
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5276 - loss: 0.8504
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5274 - loss: 0.8508
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5271 - loss: 0.8512
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5269 - loss: 0.8516
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5267 - loss: 0.8519
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5265 - loss: 0.8522
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5264 - loss: 0.8525
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5261 - loss: 0.8528
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5259 - loss: 0.8531
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5257 - loss: 0.8534
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5255 - loss: 0.8537
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5254 - loss: 0.8540
59/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5252 - loss: 0.8543
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5250 - loss: 0.8546
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5248 - loss: 0.8548
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5247 - loss: 0.8550
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5245 - loss: 0.8552
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5244 - loss: 0.8554
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5242 - loss: 0.8556
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5240 - loss: 0.8558
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5239 - loss: 0.8560
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5237 - loss: 0.8562
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5235 - loss: 0.8564
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5233 - loss: 0.8565
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5231 - loss: 0.8567
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5229 - loss: 0.8569
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5227 - loss: 0.8570
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5225 - loss: 0.8572
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5223 - loss: 0.8574
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5221 - loss: 0.8575
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5219 - loss: 0.8577
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5216 - loss: 0.8579
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5214 - loss: 0.8581
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5211 - loss: 0.8583
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5209 - loss: 0.8584
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5207 - loss: 0.8586
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5204 - loss: 0.8588
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5202 - loss: 0.8590
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5200 - loss: 0.8591
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5199 - loss: 0.8593
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5197 - loss: 0.8595
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5195 - loss: 0.8597
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5193 - loss: 0.8598
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5191 - loss: 0.8600
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5189 - loss: 0.8602
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5187 - loss: 0.8603
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 138ms/step - accuracy: 0.5015 - loss: 0.8754 - val_accuracy: 0.3713 - val_loss: 0.9831
Epoch 6/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.5000 - loss: 0.8555
 2/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4961 - loss: 0.8398
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.4991 - loss: 0.8451
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5062 - loss: 0.8394
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5112 - loss: 0.8364
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5093 - loss: 0.8383
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5077 - loss: 0.8403
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5062 - loss: 0.8413
 9/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5059 - loss: 0.8424
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 120ms/step - accuracy: 0.5071 - loss: 0.8425
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5073 - loss: 0.8431
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5083 - loss: 0.8431
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5085 - loss: 0.8434
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5090 - loss: 0.8433
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5095 - loss: 0.8432
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5097 - loss: 0.8434
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5097 - loss: 0.8435
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5098 - loss: 0.8436
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5098 - loss: 0.8439
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5095 - loss: 0.8441
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5093 - loss: 0.8442
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5093 - loss: 0.8444
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5094 - loss: 0.8444
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5095 - loss: 0.8446
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5095 - loss: 0.8448
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5096 - loss: 0.8450
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5096 - loss: 0.8452
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5095 - loss: 0.8453
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5095 - loss: 0.8454
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5095 - loss: 0.8456
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5095 - loss: 0.8458
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5093 - loss: 0.8461
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5090 - loss: 0.8464
34/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5088 - loss: 0.8468
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5085 - loss: 0.8471
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5083 - loss: 0.8474
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5082 - loss: 0.8476
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5081 - loss: 0.8477
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5080 - loss: 0.8479
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5079 - loss: 0.8480
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5078 - loss: 0.8481
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5077 - loss: 0.8482
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5075 - loss: 0.8483
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5074 - loss: 0.8484
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5073 - loss: 0.8485
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5072 - loss: 0.8485
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5071 - loss: 0.8486
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5070 - loss: 0.8487
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5070 - loss: 0.8488
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5069 - loss: 0.8488
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8487
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5069 - loss: 0.8488
59/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5069 - loss: 0.8488
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5068 - loss: 0.8489
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5068 - loss: 0.8489
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5068 - loss: 0.8490
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5068 - loss: 0.8491
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5067 - loss: 0.8491
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5067 - loss: 0.8492
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5067 - loss: 0.8492
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5067 - loss: 0.8493
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5066 - loss: 0.8493
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5066 - loss: 0.8493
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5065 - loss: 0.8494
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5065 - loss: 0.8494
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5064 - loss: 0.8495
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5064 - loss: 0.8496
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5063 - loss: 0.8496
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5063 - loss: 0.8497
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5062 - loss: 0.8498
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.5062 - loss: 0.8499
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.5061 - loss: 0.8499
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.5061 - loss: 0.8500
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.5060 - loss: 0.8501
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.5060 - loss: 0.8502
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5059 - loss: 0.8502
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5059 - loss: 0.8503
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5059 - loss: 0.8504
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5059 - loss: 0.8505
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5058 - loss: 0.8506
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5058 - loss: 0.8508
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5057 - loss: 0.8509
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5057 - loss: 0.8510
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5056 - loss: 0.8512
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5056 - loss: 0.8513
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5055 - loss: 0.8514
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 138ms/step - accuracy: 0.4998 - loss: 0.8642 - val_accuracy: 0.3467 - val_loss: 0.9802
Epoch 7/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 11s 121ms/step - accuracy: 0.5781 - loss: 0.7602
 2/92 ━━━━━━━━━━━━━━━━━━━━ 11s 124ms/step - accuracy: 0.5352 - loss: 0.7900
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.5148 - loss: 0.8007
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.5091 - loss: 0.8058
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.5092 - loss: 0.8073
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5102 - loss: 0.8085
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5123 - loss: 0.8090
 8/92 ━━━━━━━━━━━━━━━━━━━━ 9s 116ms/step - accuracy: 0.5125 - loss: 0.8096
 9/92 ━━━━━━━━━━━━━━━━━━━━ 9s 117ms/step - accuracy: 0.5131 - loss: 0.8110
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5126 - loss: 0.8133
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5119 - loss: 0.8157
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5110 - loss: 0.8178
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5099 - loss: 0.8197
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5095 - loss: 0.8208
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5088 - loss: 0.8220
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 119ms/step - accuracy: 0.5085 - loss: 0.8233
17/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.5085 - loss: 0.8240
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 119ms/step - accuracy: 0.5085 - loss: 0.8246
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5086 - loss: 0.8251
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5087 - loss: 0.8255
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5090 - loss: 0.8259
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5091 - loss: 0.8264
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5090 - loss: 0.8269
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5089 - loss: 0.8276
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5091 - loss: 0.8280
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5090 - loss: 0.8285
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5090 - loss: 0.8289
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5090 - loss: 0.8294
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5090 - loss: 0.8298
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5090 - loss: 0.8303
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5089 - loss: 0.8307
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5088 - loss: 0.8311
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 120ms/step - accuracy: 0.5088 - loss: 0.8314
34/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5087 - loss: 0.8317
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5086 - loss: 0.8321
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5085 - loss: 0.8325
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5084 - loss: 0.8327
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5084 - loss: 0.8330
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5083 - loss: 0.8332
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5083 - loss: 0.8333
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5083 - loss: 0.8334
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5084 - loss: 0.8335
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5084 - loss: 0.8336
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5085 - loss: 0.8337
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5085 - loss: 0.8338
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5085 - loss: 0.8339
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5085 - loss: 0.8341
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5085 - loss: 0.8343
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5085 - loss: 0.8345
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5085 - loss: 0.8346
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.5085 - loss: 0.8347
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5085 - loss: 0.8347
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5086 - loss: 0.8348
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.5086 - loss: 0.8349
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.5086 - loss: 0.8349
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.5087 - loss: 0.8349
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.5087 - loss: 0.8350
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 120ms/step - accuracy: 0.5087 - loss: 0.8350
59/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5087 - loss: 0.8350
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5088 - loss: 0.8351
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5088 - loss: 0.8351
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5087 - loss: 0.8351
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5087 - loss: 0.8352
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5087 - loss: 0.8352
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5087 - loss: 0.8352
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5087 - loss: 0.8352
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 120ms/step - accuracy: 0.5086 - loss: 0.8352
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5087 - loss: 0.8351
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.5086 - loss: 0.8351
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.5086 - loss: 0.8351
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5086 - loss: 0.8351
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5085 - loss: 0.8350
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5085 - loss: 0.8350
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5085 - loss: 0.8350
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5084 - loss: 0.8350
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5084 - loss: 0.8350
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5083 - loss: 0.8350
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5083 - loss: 0.8350
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5082 - loss: 0.8350
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5081 - loss: 0.8351
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5080 - loss: 0.8351
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5080 - loss: 0.8351
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5079 - loss: 0.8351
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5078 - loss: 0.8352
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 120ms/step - accuracy: 0.5078 - loss: 0.8352
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5077 - loss: 0.8353
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 137ms/step - accuracy: 0.5017 - loss: 0.8401 - val_accuracy: 0.3433 - val_loss: 1.0066
Epoch 8/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5156 - loss: 0.8631
 2/92 ━━━━━━━━━━━━━━━━━━━━ 11s 129ms/step - accuracy: 0.4961 - loss: 0.8516
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 123ms/step - accuracy: 0.4835 - loss: 0.8489
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.4769 - loss: 0.8490
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 123ms/step - accuracy: 0.4778 - loss: 0.8468
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.4780 - loss: 0.8463
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.4783 - loss: 0.8463
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.4778 - loss: 0.8460
 9/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.4780 - loss: 0.8454
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4783 - loss: 0.8451
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4788 - loss: 0.8446
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4795 - loss: 0.8437
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4804 - loss: 0.8428
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4812 - loss: 0.8420
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4823 - loss: 0.8410
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4833 - loss: 0.8402
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.4839 - loss: 0.8396
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4848 - loss: 0.8388
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4856 - loss: 0.8379
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4866 - loss: 0.8369
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4873 - loss: 0.8361
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4879 - loss: 0.8355
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4882 - loss: 0.8352
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4885 - loss: 0.8349
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.4890 - loss: 0.8344
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4895 - loss: 0.8340
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4901 - loss: 0.8336
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4905 - loss: 0.8331
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4908 - loss: 0.8326
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4913 - loss: 0.8323
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4917 - loss: 0.8320
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4920 - loss: 0.8318
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4923 - loss: 0.8316
34/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.4926 - loss: 0.8313
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4928 - loss: 0.8311
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4930 - loss: 0.8309
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4932 - loss: 0.8307
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4934 - loss: 0.8305
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4936 - loss: 0.8304
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4938 - loss: 0.8302
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4940 - loss: 0.8301
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.4941 - loss: 0.8299
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4943 - loss: 0.8297
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4946 - loss: 0.8295
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4949 - loss: 0.8294
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4952 - loss: 0.8291
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4955 - loss: 0.8290
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4957 - loss: 0.8289
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4959 - loss: 0.8287
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.4961 - loss: 0.8286
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4963 - loss: 0.8285
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4965 - loss: 0.8284
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4967 - loss: 0.8283
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4969 - loss: 0.8281
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4970 - loss: 0.8281
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4971 - loss: 0.8280
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4972 - loss: 0.8280
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.4973 - loss: 0.8279
59/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4973 - loss: 0.8279
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4974 - loss: 0.8279
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4974 - loss: 0.8279
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4975 - loss: 0.8278
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4975 - loss: 0.8278
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4975 - loss: 0.8278
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4975 - loss: 0.8278
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4976 - loss: 0.8277
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.4976 - loss: 0.8277
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4976 - loss: 0.8277
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4977 - loss: 0.8277
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4977 - loss: 0.8277
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4977 - loss: 0.8276
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4977 - loss: 0.8277
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4977 - loss: 0.8277
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.4978 - loss: 0.8277
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 120ms/step - accuracy: 0.4978 - loss: 0.8278
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4978 - loss: 0.8278
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 120ms/step - accuracy: 0.4978 - loss: 0.8278
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4978 - loss: 0.8279
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4978 - loss: 0.8279
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4978 - loss: 0.8279
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4978 - loss: 0.8279
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4979 - loss: 0.8279
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.4979 - loss: 0.8280
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4980 - loss: 0.8280
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4980 - loss: 0.8280
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4980 - loss: 0.8280
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4981 - loss: 0.8281
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4981 - loss: 0.8281
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4981 - loss: 0.8281
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4981 - loss: 0.8281
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4981 - loss: 0.8281
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.4981 - loss: 0.8281
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 138ms/step - accuracy: 0.5005 - loss: 0.8282 - val_accuracy: 0.3440 - val_loss: 1.0181
Epoch 9/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 11s 122ms/step - accuracy: 0.6094 - loss: 0.7861
 2/92 ━━━━━━━━━━━━━━━━━━━━ 10s 119ms/step - accuracy: 0.5977 - loss: 0.7713
 3/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5859 - loss: 0.7708
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5752 - loss: 0.7738
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5702 - loss: 0.7766
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 120ms/step - accuracy: 0.5650 - loss: 0.7816
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 121ms/step - accuracy: 0.5627 - loss: 0.7830
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.5610 - loss: 0.7846
 9/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.5602 - loss: 0.7853
10/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.5595 - loss: 0.7852
11/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.5573 - loss: 0.7860
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 122ms/step - accuracy: 0.5553 - loss: 0.7868
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5541 - loss: 0.7875
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5531 - loss: 0.7881
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 121ms/step - accuracy: 0.5520 - loss: 0.7885
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 124ms/step - accuracy: 0.5508 - loss: 0.7888
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 127ms/step - accuracy: 0.5497 - loss: 0.7889
18/92 ━━━━━━━━━━━━━━━━━━━━ 9s 128ms/step - accuracy: 0.5484 - loss: 0.7891
19/92 ━━━━━━━━━━━━━━━━━━━━ 9s 127ms/step - accuracy: 0.5474 - loss: 0.7894
20/92 ━━━━━━━━━━━━━━━━━━━━ 9s 127ms/step - accuracy: 0.5464 - loss: 0.7896
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 126ms/step - accuracy: 0.5456 - loss: 0.7896
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 125ms/step - accuracy: 0.5448 - loss: 0.7896
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 125ms/step - accuracy: 0.5440 - loss: 0.7896
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 125ms/step - accuracy: 0.5432 - loss: 0.7896
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 124ms/step - accuracy: 0.5424 - loss: 0.7897
26/92 ━━━━━━━━━━━━━━━━━━━━ 8s 124ms/step - accuracy: 0.5415 - loss: 0.7899
27/92 ━━━━━━━━━━━━━━━━━━━━ 8s 124ms/step - accuracy: 0.5406 - loss: 0.7902
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 124ms/step - accuracy: 0.5398 - loss: 0.7904
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 124ms/step - accuracy: 0.5391 - loss: 0.7908
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 124ms/step - accuracy: 0.5384 - loss: 0.7911
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 123ms/step - accuracy: 0.5378 - loss: 0.7914
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 124ms/step - accuracy: 0.5372 - loss: 0.7916
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 123ms/step - accuracy: 0.5367 - loss: 0.7918
34/92 ━━━━━━━━━━━━━━━━━━━━ 7s 123ms/step - accuracy: 0.5362 - loss: 0.7920
35/92 ━━━━━━━━━━━━━━━━━━━━ 7s 123ms/step - accuracy: 0.5356 - loss: 0.7922
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5351 - loss: 0.7924
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5346 - loss: 0.7926
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5342 - loss: 0.7929
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5338 - loss: 0.7931
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5333 - loss: 0.7933
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5329 - loss: 0.7935
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5325 - loss: 0.7937
43/92 ━━━━━━━━━━━━━━━━━━━━ 6s 123ms/step - accuracy: 0.5321 - loss: 0.7940
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 123ms/step - accuracy: 0.5317 - loss: 0.7942
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 123ms/step - accuracy: 0.5313 - loss: 0.7945
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 123ms/step - accuracy: 0.5309 - loss: 0.7948
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 123ms/step - accuracy: 0.5304 - loss: 0.7951
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5300 - loss: 0.7954
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5296 - loss: 0.7957
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5292 - loss: 0.7960
51/92 ━━━━━━━━━━━━━━━━━━━━ 5s 122ms/step - accuracy: 0.5288 - loss: 0.7963
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5285 - loss: 0.7965
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5281 - loss: 0.7967
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5278 - loss: 0.7970
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5275 - loss: 0.7971
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5272 - loss: 0.7973
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5269 - loss: 0.7974
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5267 - loss: 0.7976
59/92 ━━━━━━━━━━━━━━━━━━━━ 4s 122ms/step - accuracy: 0.5265 - loss: 0.7977
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5263 - loss: 0.7978
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5260 - loss: 0.7979
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5258 - loss: 0.7981
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5257 - loss: 0.7982
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5255 - loss: 0.7982
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5253 - loss: 0.7984
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5252 - loss: 0.7984
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 122ms/step - accuracy: 0.5251 - loss: 0.7985
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5250 - loss: 0.7986
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5248 - loss: 0.7988
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5247 - loss: 0.7988
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5246 - loss: 0.7989
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5245 - loss: 0.7990
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5244 - loss: 0.7991
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5243 - loss: 0.7993
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 122ms/step - accuracy: 0.5242 - loss: 0.7994
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5241 - loss: 0.7995
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5239 - loss: 0.7996
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5238 - loss: 0.7997
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5237 - loss: 0.7998
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5236 - loss: 0.8000
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5235 - loss: 0.8001
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5233 - loss: 0.8003
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 122ms/step - accuracy: 0.5232 - loss: 0.8004
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.5231 - loss: 0.8006
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5230 - loss: 0.8007
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5228 - loss: 0.8009
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 122ms/step - accuracy: 0.5227 - loss: 0.8010
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5226 - loss: 0.8011
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5225 - loss: 0.8013
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5225 - loss: 0.8014
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5224 - loss: 0.8015
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5223 - loss: 0.8017
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 138ms/step - accuracy: 0.5148 - loss: 0.8148 - val_accuracy: 0.3208 - val_loss: 1.0333
Epoch 10/10
 1/92 ━━━━━━━━━━━━━━━━━━━━ 11s 130ms/step - accuracy: 0.5000 - loss: 0.7813
 2/92 ━━━━━━━━━━━━━━━━━━━━ 11s 128ms/step - accuracy: 0.4922 - loss: 0.8061
 3/92 ━━━━━━━━━━━━━━━━━━━━ 11s 124ms/step - accuracy: 0.4844 - loss: 0.8123
 4/92 ━━━━━━━━━━━━━━━━━━━━ 10s 122ms/step - accuracy: 0.4834 - loss: 0.8139
 5/92 ━━━━━━━━━━━━━━━━━━━━ 10s 126ms/step - accuracy: 0.4855 - loss: 0.8159
 6/92 ━━━━━━━━━━━━━━━━━━━━ 10s 125ms/step - accuracy: 0.4888 - loss: 0.8166
 7/92 ━━━━━━━━━━━━━━━━━━━━ 10s 124ms/step - accuracy: 0.4900 - loss: 0.8164
 8/92 ━━━━━━━━━━━━━━━━━━━━ 10s 126ms/step - accuracy: 0.4906 - loss: 0.8158
 9/92 ━━━━━━━━━━━━━━━━━━━━ 10s 126ms/step - accuracy: 0.4914 - loss: 0.8153
10/92 ━━━━━━━━━━━━━━━━━━━━ 10s 126ms/step - accuracy: 0.4918 - loss: 0.8153
11/92 ━━━━━━━━━━━━━━━━━━━━ 10s 125ms/step - accuracy: 0.4928 - loss: 0.8148
12/92 ━━━━━━━━━━━━━━━━━━━━ 9s 125ms/step - accuracy: 0.4937 - loss: 0.8146
13/92 ━━━━━━━━━━━━━━━━━━━━ 9s 124ms/step - accuracy: 0.4948 - loss: 0.8141
14/92 ━━━━━━━━━━━━━━━━━━━━ 9s 124ms/step - accuracy: 0.4955 - loss: 0.8136
15/92 ━━━━━━━━━━━━━━━━━━━━ 9s 123ms/step - accuracy: 0.4966 - loss: 0.8127
16/92 ━━━━━━━━━━━━━━━━━━━━ 9s 123ms/step - accuracy: 0.4979 - loss: 0.8115
17/92 ━━━━━━━━━━━━━━━━━━━━ 9s 123ms/step - accuracy: 0.4990 - loss: 0.8105
18/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5001 - loss: 0.8096
19/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5010 - loss: 0.8091
20/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5017 - loss: 0.8086
21/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5023 - loss: 0.8081
22/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5029 - loss: 0.8076
23/92 ━━━━━━━━━━━━━━━━━━━━ 8s 121ms/step - accuracy: 0.5036 - loss: 0.8071
24/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5042 - loss: 0.8066
25/92 ━━━━━━━━━━━━━━━━━━━━ 8s 120ms/step - accuracy: 0.5050 - loss: 0.8060
26/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5054 - loss: 0.8056
27/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5058 - loss: 0.8051
28/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5061 - loss: 0.8047
29/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5065 - loss: 0.8044
30/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5069 - loss: 0.8041
31/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5071 - loss: 0.8039
32/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5073 - loss: 0.8036
33/92 ━━━━━━━━━━━━━━━━━━━━ 7s 121ms/step - accuracy: 0.5075 - loss: 0.8033
34/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5078 - loss: 0.8031
35/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5080 - loss: 0.8028
36/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5083 - loss: 0.8026
37/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5086 - loss: 0.8023
38/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5088 - loss: 0.8022
39/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5090 - loss: 0.8020
40/92 ━━━━━━━━━━━━━━━━━━━━ 6s 121ms/step - accuracy: 0.5093 - loss: 0.8018
41/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5095 - loss: 0.8016
42/92 ━━━━━━━━━━━━━━━━━━━━ 6s 120ms/step - accuracy: 0.5096 - loss: 0.8014
43/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5098 - loss: 0.8013
44/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5099 - loss: 0.8010
45/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5101 - loss: 0.8009
46/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5102 - loss: 0.8007
47/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5102 - loss: 0.8006
48/92 ━━━━━━━━━━━━━━━━━━━━ 5s 120ms/step - accuracy: 0.5103 - loss: 0.8004
49/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5104 - loss: 0.8003
50/92 ━━━━━━━━━━━━━━━━━━━━ 5s 121ms/step - accuracy: 0.5105 - loss: 0.8002
51/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5106 - loss: 0.8001
52/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5106 - loss: 0.8000
53/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5107 - loss: 0.8000
54/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5108 - loss: 0.7999
55/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5108 - loss: 0.7999
56/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5108 - loss: 0.7999
57/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5108 - loss: 0.7999
58/92 ━━━━━━━━━━━━━━━━━━━━ 4s 121ms/step - accuracy: 0.5108 - loss: 0.7998
59/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5109 - loss: 0.7998
60/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5109 - loss: 0.7998
61/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5109 - loss: 0.7998
62/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5110 - loss: 0.7998
63/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5110 - loss: 0.7998
64/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5111 - loss: 0.7998
65/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5111 - loss: 0.7998
66/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5112 - loss: 0.7998
67/92 ━━━━━━━━━━━━━━━━━━━━ 3s 121ms/step - accuracy: 0.5113 - loss: 0.7998
68/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5114 - loss: 0.7997
69/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5115 - loss: 0.7997
70/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5116 - loss: 0.7997
71/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5116 - loss: 0.7996
72/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5117 - loss: 0.7996
73/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5118 - loss: 0.7995
74/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5119 - loss: 0.7995
75/92 ━━━━━━━━━━━━━━━━━━━━ 2s 121ms/step - accuracy: 0.5120 - loss: 0.7995
76/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5120 - loss: 0.7995
77/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5121 - loss: 0.7994
78/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5121 - loss: 0.7994
79/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5121 - loss: 0.7994
80/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5122 - loss: 0.7994
81/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5122 - loss: 0.7994
82/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5122 - loss: 0.7993
83/92 ━━━━━━━━━━━━━━━━━━━━ 1s 121ms/step - accuracy: 0.5122 - loss: 0.7993
84/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5122 - loss: 0.7993
85/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7993
86/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7994
87/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7994
88/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7994
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7994
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7994
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7995
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 121ms/step - accuracy: 0.5123 - loss: 0.7995
92/92 ━━━━━━━━━━━━━━━━━━━━ 13s 138ms/step - accuracy: 0.5116 - loss: 0.8018 - val_accuracy: 0.3297 - val_loss: 1.0562

Hybrid Model Results:
  Validation Loss: 1.0562
  Validation Accuracy: 0.3297

Hybrid Model Characteristics:

• Trains both custom and pre-trained pathways

• Custom layers can learn flower-specific patterns (petal textures, colors)

• Pre-trained features provide robust general visual understanding

• More trainable parameters than frozen transfer learning

• Useful when your domain has unique characteristics not in ImageNet

Fine-tuning

Fine-tuning is the second transfer learning strategy. After training the new classification head with a frozen backbone, we:

1. Unfreeze some or all of the backbone layers

2. Continue training with a much lower learning rate

3. Allow the pre-trained weights to adapt slightly to our specific task

Why Fine-tune?

• Better adaptation: Backbone features adjust to our specific dataset

• Higher accuracy: Often achieves better results than frozen features

• Task-specific features: Top layers learn patterns specific to flowers

Risks and Precautions

• Catastrophic forgetting: High learning rate can destroy pre-trained knowledge

• Overfitting: More trainable parameters increase overfitting risk

• Solution: Use very small learning rate (10-100× smaller than initial training)

Unfreezing the Backbone

We can unfreeze:

• All layers: Maximum flexibility

• Top N layers: Keeps early generic features frozen

Let’s unfreeze the entire backbone:

# Unfreeze the base model - allows weights to update during training
base_model.trainable = True

print(f'Number of layers in base model: {len(base_model.layers)}')
print(f'Trainable layers: {sum([1 for layer in base_model.layers if layer.trainable])}')

Number of layers in base model: 154
Trainable layers: 154

Recompiling with Lower Learning Rate

Critical: Use a much smaller learning rate for fine-tuning to avoid destroying pre-trained knowledge.

# Compile with very low learning rate (100× smaller than default)
# This prevents catastrophic forgetting of ImageNet features
model_transfer.compile(
    optimizer=keras.optimizers.Adam(learning_rate=1e-5),  # Default is 1e-3
    loss=tf.keras.losses.SparseCategoricalCrossentropy(),
    metrics=['accuracy']
)

print("Model recompiled with learning rate: 1e-5")

Model recompiled with learning rate: 1e-5

Fine-tuning Training

Continue training for a few more epochs.

epochs_finetune = 5
total_epochs = epochs_initial + epochs_finetune

# Continue training from where we left off
history_finetune = model_transfer.fit(
    train_ds,
    validation_data=val_ds,
    epochs=total_epochs,
    initial_epoch=history_transfer.epoch[-1]
)

Epoch 10/15
 1/92 ━━━━━━━━━━━━━━━━━━━━ 13:18 9s/step - accuracy: 0.4375 - loss: 1.4773
 2/92 ━━━━━━━━━━━━━━━━━━━━ 22s 250ms/step - accuracy: 0.4258 - loss: 1.5109
 3/92 ━━━━━━━━━━━━━━━━━━━━ 22s 254ms/step - accuracy: 0.4297 - loss: 1.4957
 4/92 ━━━━━━━━━━━━━━━━━━━━ 22s 253ms/step - accuracy: 0.4307 - loss: 1.4833
 5/92 ━━━━━━━━━━━━━━━━━━━━ 22s 258ms/step - accuracy: 0.4308 - loss: 1.4694
 6/92 ━━━━━━━━━━━━━━━━━━━━ 22s 259ms/step - accuracy: 0.4306 - loss: 1.4629
 7/92 ━━━━━━━━━━━━━━━━━━━━ 22s 260ms/step - accuracy: 0.4316 - loss: 1.4562
 8/92 ━━━━━━━━━━━━━━━━━━━━ 21s 260ms/step - accuracy: 0.4301 - loss: 1.4505
 9/92 ━━━━━━━━━━━━━━━━━━━━ 21s 260ms/step - accuracy: 0.4277 - loss: 1.4487
10/92 ━━━━━━━━━━━━━━━━━━━━ 21s 260ms/step - accuracy: 0.4252 - loss: 1.4472
11/92 ━━━━━━━━━━━━━━━━━━━━ 21s 261ms/step - accuracy: 0.4227 - loss: 1.4460
12/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4207 - loss: 1.4452
13/92 ━━━━━━━━━━━━━━━━━━━━ 20s 260ms/step - accuracy: 0.4192 - loss: 1.4444
14/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4184 - loss: 1.4425
15/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4179 - loss: 1.4403
16/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4175 - loss: 1.4378
17/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4170 - loss: 1.4358
18/92 ━━━━━━━━━━━━━━━━━━━━ 19s 262ms/step - accuracy: 0.4166 - loss: 1.4335
19/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4162 - loss: 1.4315
20/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4158 - loss: 1.4294
21/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4157 - loss: 1.4274
22/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4154 - loss: 1.4257
23/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4152 - loss: 1.4244
24/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4150 - loss: 1.4231
25/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4147 - loss: 1.4220
26/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4145 - loss: 1.4209
27/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.4144 - loss: 1.4195
28/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.4143 - loss: 1.4184
29/92 ━━━━━━━━━━━━━━━━━━━━ 16s 261ms/step - accuracy: 0.4141 - loss: 1.4173
30/92 ━━━━━━━━━━━━━━━━━━━━ 16s 261ms/step - accuracy: 0.4141 - loss: 1.4160
31/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4142 - loss: 1.4150
32/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4141 - loss: 1.4143
33/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4140 - loss: 1.4136
34/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4139 - loss: 1.4130
35/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4139 - loss: 1.4123
36/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4138 - loss: 1.4117
37/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4136 - loss: 1.4111
38/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4134 - loss: 1.4106
39/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4133 - loss: 1.4100
40/92 ━━━━━━━━━━━━━━━━━━━━ 13s 260ms/step - accuracy: 0.4133 - loss: 1.4094
41/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4132 - loss: 1.4087
42/92 ━━━━━━━━━━━━━━━━━━━━ 13s 260ms/step - accuracy: 0.4131 - loss: 1.4081
43/92 ━━━━━━━━━━━━━━━━━━━━ 12s 260ms/step - accuracy: 0.4130 - loss: 1.4074
44/92 ━━━━━━━━━━━━━━━━━━━━ 12s 260ms/step - accuracy: 0.4130 - loss: 1.4067
45/92 ━━━━━━━━━━━━━━━━━━━━ 12s 260ms/step - accuracy: 0.4130 - loss: 1.4060
46/92 ━━━━━━━━━━━━━━━━━━━━ 11s 260ms/step - accuracy: 0.4130 - loss: 1.4053
47/92 ━━━━━━━━━━━━━━━━━━━━ 11s 260ms/step - accuracy: 0.4130 - loss: 1.4047
48/92 ━━━━━━━━━━━━━━━━━━━━ 11s 260ms/step - accuracy: 0.4131 - loss: 1.4041
49/92 ━━━━━━━━━━━━━━━━━━━━ 11s 260ms/step - accuracy: 0.4131 - loss: 1.4035
50/92 ━━━━━━━━━━━━━━━━━━━━ 10s 260ms/step - accuracy: 0.4131 - loss: 1.4030
51/92 ━━━━━━━━━━━━━━━━━━━━ 10s 260ms/step - accuracy: 0.4132 - loss: 1.4025
52/92 ━━━━━━━━━━━━━━━━━━━━ 10s 260ms/step - accuracy: 0.4132 - loss: 1.4019
53/92 ━━━━━━━━━━━━━━━━━━━━ 10s 260ms/step - accuracy: 0.4133 - loss: 1.4014
54/92 ━━━━━━━━━━━━━━━━━━━━ 9s 260ms/step - accuracy: 0.4133 - loss: 1.4008
55/92 ━━━━━━━━━━━━━━━━━━━━ 9s 260ms/step - accuracy: 0.4134 - loss: 1.4002
56/92 ━━━━━━━━━━━━━━━━━━━━ 9s 260ms/step - accuracy: 0.4134 - loss: 1.3996
57/92 ━━━━━━━━━━━━━━━━━━━━ 9s 260ms/step - accuracy: 0.4135 - loss: 1.3990
58/92 ━━━━━━━━━━━━━━━━━━━━ 8s 260ms/step - accuracy: 0.4135 - loss: 1.3985
59/92 ━━━━━━━━━━━━━━━━━━━━ 8s 260ms/step - accuracy: 0.4135 - loss: 1.3980
60/92 ━━━━━━━━━━━━━━━━━━━━ 8s 260ms/step - accuracy: 0.4135 - loss: 1.3975
61/92 ━━━━━━━━━━━━━━━━━━━━ 8s 260ms/step - accuracy: 0.4136 - loss: 1.3970
62/92 ━━━━━━━━━━━━━━━━━━━━ 7s 260ms/step - accuracy: 0.4136 - loss: 1.3966
63/92 ━━━━━━━━━━━━━━━━━━━━ 7s 260ms/step - accuracy: 0.4136 - loss: 1.3962
64/92 ━━━━━━━━━━━━━━━━━━━━ 7s 260ms/step - accuracy: 0.4136 - loss: 1.3957
65/92 ━━━━━━━━━━━━━━━━━━━━ 7s 260ms/step - accuracy: 0.4137 - loss: 1.3953
66/92 ━━━━━━━━━━━━━━━━━━━━ 6s 260ms/step - accuracy: 0.4137 - loss: 1.3949
67/92 ━━━━━━━━━━━━━━━━━━━━ 6s 260ms/step - accuracy: 0.4137 - loss: 1.3946
68/92 ━━━━━━━━━━━━━━━━━━━━ 6s 260ms/step - accuracy: 0.4137 - loss: 1.3942
69/92 ━━━━━━━━━━━━━━━━━━━━ 5s 260ms/step - accuracy: 0.4138 - loss: 1.3938
70/92 ━━━━━━━━━━━━━━━━━━━━ 5s 261ms/step - accuracy: 0.4138 - loss: 1.3935
71/92 ━━━━━━━━━━━━━━━━━━━━ 5s 260ms/step - accuracy: 0.4138 - loss: 1.3931
72/92 ━━━━━━━━━━━━━━━━━━━━ 5s 260ms/step - accuracy: 0.4138 - loss: 1.3928
73/92 ━━━━━━━━━━━━━━━━━━━━ 4s 260ms/step - accuracy: 0.4138 - loss: 1.3924
74/92 ━━━━━━━━━━━━━━━━━━━━ 4s 260ms/step - accuracy: 0.4139 - loss: 1.3920
75/92 ━━━━━━━━━━━━━━━━━━━━ 4s 260ms/step - accuracy: 0.4139 - loss: 1.3916
76/92 ━━━━━━━━━━━━━━━━━━━━ 4s 260ms/step - accuracy: 0.4139 - loss: 1.3912
77/92 ━━━━━━━━━━━━━━━━━━━━ 3s 260ms/step - accuracy: 0.4139 - loss: 1.3908
78/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4140 - loss: 1.3903
79/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4140 - loss: 1.3899
80/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4140 - loss: 1.3895
81/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4141 - loss: 1.3890
82/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4141 - loss: 1.3886
83/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4141 - loss: 1.3882
84/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4142 - loss: 1.3877
85/92 ━━━━━━━━━━━━━━━━━━━━ 1s 259ms/step - accuracy: 0.4142 - loss: 1.3873
86/92 ━━━━━━━━━━━━━━━━━━━━ 1s 259ms/step - accuracy: 0.4143 - loss: 1.3868
87/92 ━━━━━━━━━━━━━━━━━━━━ 1s 259ms/step - accuracy: 0.4143 - loss: 1.3864
88/92 ━━━━━━━━━━━━━━━━━━━━ 1s 259ms/step - accuracy: 0.4144 - loss: 1.3859
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 259ms/step - accuracy: 0.4144 - loss: 1.3854
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 259ms/step - accuracy: 0.4145 - loss: 1.3850
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 259ms/step - accuracy: 0.4146 - loss: 1.3845
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 259ms/step - accuracy: 0.4146 - loss: 1.3840
92/92 ━━━━━━━━━━━━━━━━━━━━ 34s 277ms/step - accuracy: 0.4189 - loss: 1.3424 - val_accuracy: 0.3549 - val_loss: 1.1579
Epoch 11/15
 1/92 ━━━━━━━━━━━━━━━━━━━━ 24s 270ms/step - accuracy: 0.3906 - loss: 1.3795
 2/92 ━━━━━━━━━━━━━━━━━━━━ 23s 256ms/step - accuracy: 0.3945 - loss: 1.3622
 3/92 ━━━━━━━━━━━━━━━━━━━━ 22s 255ms/step - accuracy: 0.3932 - loss: 1.3509
 4/92 ━━━━━━━━━━━━━━━━━━━━ 22s 256ms/step - accuracy: 0.3955 - loss: 1.3406
 5/92 ━━━━━━━━━━━━━━━━━━━━ 24s 279ms/step - accuracy: 0.3995 - loss: 1.3279
 6/92 ━━━━━━━━━━━━━━━━━━━━ 23s 276ms/step - accuracy: 0.4046 - loss: 1.3195
 7/92 ━━━━━━━━━━━━━━━━━━━━ 23s 273ms/step - accuracy: 0.4086 - loss: 1.3132
 8/92 ━━━━━━━━━━━━━━━━━━━━ 22s 270ms/step - accuracy: 0.4108 - loss: 1.3101
 9/92 ━━━━━━━━━━━━━━━━━━━━ 21s 261ms/step - accuracy: 0.4137 - loss: 1.3062
10/92 ━━━━━━━━━━━━━━━━━━━━ 21s 260ms/step - accuracy: 0.4176 - loss: 1.3020
11/92 ━━━━━━━━━━━━━━━━━━━━ 21s 260ms/step - accuracy: 0.4204 - loss: 1.2992
12/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4223 - loss: 1.2966
13/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4241 - loss: 1.2945
14/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4262 - loss: 1.2927
15/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4278 - loss: 1.2911
16/92 ━━━━━━━━━━━━━━━━━━━━ 19s 262ms/step - accuracy: 0.4294 - loss: 1.2892
17/92 ━━━━━━━━━━━━━━━━━━━━ 19s 262ms/step - accuracy: 0.4311 - loss: 1.2868
18/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4326 - loss: 1.2842
19/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4341 - loss: 1.2813
20/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.4354 - loss: 1.2785
21/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.4367 - loss: 1.2758
22/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4379 - loss: 1.2733
23/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4390 - loss: 1.2708
24/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4401 - loss: 1.2682
25/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4412 - loss: 1.2657
26/92 ━━━━━━━━━━━━━━━━━━━━ 17s 260ms/step - accuracy: 0.4423 - loss: 1.2631
27/92 ━━━━━━━━━━━━━━━━━━━━ 16s 260ms/step - accuracy: 0.4435 - loss: 1.2605
28/92 ━━━━━━━━━━━━━━━━━━━━ 16s 261ms/step - accuracy: 0.4445 - loss: 1.2581
29/92 ━━━━━━━━━━━━━━━━━━━━ 16s 261ms/step - accuracy: 0.4455 - loss: 1.2561
30/92 ━━━━━━━━━━━━━━━━━━━━ 16s 261ms/step - accuracy: 0.4463 - loss: 1.2544
31/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4471 - loss: 1.2528
32/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4478 - loss: 1.2514
33/92 ━━━━━━━━━━━━━━━━━━━━ 15s 261ms/step - accuracy: 0.4484 - loss: 1.2500
34/92 ━━━━━━━━━━━━━━━━━━━━ 15s 260ms/step - accuracy: 0.4490 - loss: 1.2486
35/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4495 - loss: 1.2472
36/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4500 - loss: 1.2458
37/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4505 - loss: 1.2446
38/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4510 - loss: 1.2433
39/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4515 - loss: 1.2419
40/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4520 - loss: 1.2406
41/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4524 - loss: 1.2393
42/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4528 - loss: 1.2380
43/92 ━━━━━━━━━━━━━━━━━━━━ 12s 261ms/step - accuracy: 0.4532 - loss: 1.2368
44/92 ━━━━━━━━━━━━━━━━━━━━ 12s 261ms/step - accuracy: 0.4536 - loss: 1.2357
45/92 ━━━━━━━━━━━━━━━━━━━━ 12s 261ms/step - accuracy: 0.4540 - loss: 1.2347
46/92 ━━━━━━━━━━━━━━━━━━━━ 12s 261ms/step - accuracy: 0.4543 - loss: 1.2336
47/92 ━━━━━━━━━━━━━━━━━━━━ 11s 261ms/step - accuracy: 0.4547 - loss: 1.2326
48/92 ━━━━━━━━━━━━━━━━━━━━ 11s 261ms/step - accuracy: 0.4550 - loss: 1.2315
49/92 ━━━━━━━━━━━━━━━━━━━━ 11s 261ms/step - accuracy: 0.4553 - loss: 1.2304
50/92 ━━━━━━━━━━━━━━━━━━━━ 10s 261ms/step - accuracy: 0.4557 - loss: 1.2294
51/92 ━━━━━━━━━━━━━━━━━━━━ 10s 261ms/step - accuracy: 0.4560 - loss: 1.2284
52/92 ━━━━━━━━━━━━━━━━━━━━ 10s 261ms/step - accuracy: 0.4563 - loss: 1.2274
53/92 ━━━━━━━━━━━━━━━━━━━━ 10s 261ms/step - accuracy: 0.4565 - loss: 1.2265
54/92 ━━━━━━━━━━━━━━━━━━━━ 9s 261ms/step - accuracy: 0.4567 - loss: 1.2257
55/92 ━━━━━━━━━━━━━━━━━━━━ 9s 261ms/step - accuracy: 0.4570 - loss: 1.2249
56/92 ━━━━━━━━━━━━━━━━━━━━ 9s 261ms/step - accuracy: 0.4573 - loss: 1.2241
57/92 ━━━━━━━━━━━━━━━━━━━━ 9s 261ms/step - accuracy: 0.4575 - loss: 1.2235
58/92 ━━━━━━━━━━━━━━━━━━━━ 8s 261ms/step - accuracy: 0.4577 - loss: 1.2228
59/92 ━━━━━━━━━━━━━━━━━━━━ 8s 261ms/step - accuracy: 0.4579 - loss: 1.2222
60/92 ━━━━━━━━━━━━━━━━━━━━ 8s 261ms/step - accuracy: 0.4581 - loss: 1.2217
61/92 ━━━━━━━━━━━━━━━━━━━━ 8s 261ms/step - accuracy: 0.4583 - loss: 1.2211
62/92 ━━━━━━━━━━━━━━━━━━━━ 7s 261ms/step - accuracy: 0.4585 - loss: 1.2205
63/92 ━━━━━━━━━━━━━━━━━━━━ 7s 261ms/step - accuracy: 0.4587 - loss: 1.2200
64/92 ━━━━━━━━━━━━━━━━━━━━ 7s 261ms/step - accuracy: 0.4588 - loss: 1.2195
65/92 ━━━━━━━━━━━━━━━━━━━━ 7s 261ms/step - accuracy: 0.4590 - loss: 1.2190
66/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.4591 - loss: 1.2186
67/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.4593 - loss: 1.2181
68/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.4594 - loss: 1.2177
69/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.4596 - loss: 1.2172
70/92 ━━━━━━━━━━━━━━━━━━━━ 5s 262ms/step - accuracy: 0.4597 - loss: 1.2168
71/92 ━━━━━━━━━━━━━━━━━━━━ 5s 262ms/step - accuracy: 0.4598 - loss: 1.2165
72/92 ━━━━━━━━━━━━━━━━━━━━ 5s 262ms/step - accuracy: 0.4599 - loss: 1.2161
73/92 ━━━━━━━━━━━━━━━━━━━━ 4s 262ms/step - accuracy: 0.4601 - loss: 1.2157
74/92 ━━━━━━━━━━━━━━━━━━━━ 4s 262ms/step - accuracy: 0.4602 - loss: 1.2154
75/92 ━━━━━━━━━━━━━━━━━━━━ 4s 262ms/step - accuracy: 0.4603 - loss: 1.2150
76/92 ━━━━━━━━━━━━━━━━━━━━ 4s 262ms/step - accuracy: 0.4604 - loss: 1.2147
77/92 ━━━━━━━━━━━━━━━━━━━━ 3s 262ms/step - accuracy: 0.4604 - loss: 1.2143
78/92 ━━━━━━━━━━━━━━━━━━━━ 3s 262ms/step - accuracy: 0.4605 - loss: 1.2139
79/92 ━━━━━━━━━━━━━━━━━━━━ 3s 262ms/step - accuracy: 0.4606 - loss: 1.2136
80/92 ━━━━━━━━━━━━━━━━━━━━ 3s 262ms/step - accuracy: 0.4607 - loss: 1.2132
81/92 ━━━━━━━━━━━━━━━━━━━━ 2s 263ms/step - accuracy: 0.4608 - loss: 1.2128
82/92 ━━━━━━━━━━━━━━━━━━━━ 2s 263ms/step - accuracy: 0.4609 - loss: 1.2125
83/92 ━━━━━━━━━━━━━━━━━━━━ 2s 263ms/step - accuracy: 0.4610 - loss: 1.2122
84/92 ━━━━━━━━━━━━━━━━━━━━ 2s 263ms/step - accuracy: 0.4611 - loss: 1.2118
85/92 ━━━━━━━━━━━━━━━━━━━━ 1s 263ms/step - accuracy: 0.4611 - loss: 1.2115
86/92 ━━━━━━━━━━━━━━━━━━━━ 1s 263ms/step - accuracy: 0.4612 - loss: 1.2112
87/92 ━━━━━━━━━━━━━━━━━━━━ 1s 263ms/step - accuracy: 0.4613 - loss: 1.2109
88/92 ━━━━━━━━━━━━━━━━━━━━ 1s 263ms/step - accuracy: 0.4614 - loss: 1.2106
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 263ms/step - accuracy: 0.4614 - loss: 1.2103
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 263ms/step - accuracy: 0.4615 - loss: 1.2100
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 263ms/step - accuracy: 0.4616 - loss: 1.2096
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 262ms/step - accuracy: 0.4616 - loss: 1.2093
92/92 ━━━━━━━━━━━━━━━━━━━━ 25s 275ms/step - accuracy: 0.4671 - loss: 1.1804 - val_accuracy: 0.3522 - val_loss: 1.1616
Epoch 12/15
 1/92 ━━━━━━━━━━━━━━━━━━━━ 23s 259ms/step - accuracy: 0.4062 - loss: 1.3084
 2/92 ━━━━━━━━━━━━━━━━━━━━ 22s 249ms/step - accuracy: 0.4219 - loss: 1.2595
 3/92 ━━━━━━━━━━━━━━━━━━━━ 22s 254ms/step - accuracy: 0.4392 - loss: 1.2329
 4/92 ━━━━━━━━━━━━━━━━━━━━ 22s 256ms/step - accuracy: 0.4486 - loss: 1.2155
 5/92 ━━━━━━━━━━━━━━━━━━━━ 22s 255ms/step - accuracy: 0.4582 - loss: 1.1988
 6/92 ━━━━━━━━━━━━━━━━━━━━ 21s 254ms/step - accuracy: 0.4630 - loss: 1.1878
 7/92 ━━━━━━━━━━━━━━━━━━━━ 21s 254ms/step - accuracy: 0.4667 - loss: 1.1803
 8/92 ━━━━━━━━━━━━━━━━━━━━ 21s 257ms/step - accuracy: 0.4699 - loss: 1.1744
 9/92 ━━━━━━━━━━━━━━━━━━━━ 21s 257ms/step - accuracy: 0.4707 - loss: 1.1718
10/92 ━━━━━━━━━━━━━━━━━━━━ 21s 257ms/step - accuracy: 0.4719 - loss: 1.1685
11/92 ━━━━━━━━━━━━━━━━━━━━ 20s 257ms/step - accuracy: 0.4726 - loss: 1.1656
12/92 ━━━━━━━━━━━━━━━━━━━━ 20s 257ms/step - accuracy: 0.4728 - loss: 1.1626
13/92 ━━━━━━━━━━━━━━━━━━━━ 20s 257ms/step - accuracy: 0.4731 - loss: 1.1596
14/92 ━━━━━━━━━━━━━━━━━━━━ 20s 259ms/step - accuracy: 0.4734 - loss: 1.1572
15/92 ━━━━━━━━━━━━━━━━━━━━ 20s 261ms/step - accuracy: 0.4742 - loss: 1.1548
16/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4749 - loss: 1.1526
17/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4752 - loss: 1.1507
18/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4752 - loss: 1.1499
19/92 ━━━━━━━━━━━━━━━━━━━━ 19s 261ms/step - accuracy: 0.4753 - loss: 1.1487
20/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4756 - loss: 1.1476
21/92 ━━━━━━━━━━━━━━━━━━━━ 18s 261ms/step - accuracy: 0.4755 - loss: 1.1467
22/92 ━━━━━━━━━━━━━━━━━━━━ 18s 260ms/step - accuracy: 0.4757 - loss: 1.1456
23/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4760 - loss: 1.1441
24/92 ━━━━━━━━━━━━━━━━━━━━ 17s 261ms/step - accuracy: 0.4762 - loss: 1.1429
25/92 ━━━━━━━━━━━━━━━━━━━━ 17s 260ms/step - accuracy: 0.4766 - loss: 1.1415
26/92 ━━━━━━━━━━━━━━━━━━━━ 17s 260ms/step - accuracy: 0.4770 - loss: 1.1403
27/92 ━━━━━━━━━━━━━━━━━━━━ 16s 260ms/step - accuracy: 0.4774 - loss: 1.1390
28/92 ━━━━━━━━━━━━━━━━━━━━ 16s 260ms/step - accuracy: 0.4778 - loss: 1.1377
29/92 ━━━━━━━━━━━━━━━━━━━━ 16s 260ms/step - accuracy: 0.4781 - loss: 1.1367
30/92 ━━━━━━━━━━━━━━━━━━━━ 16s 260ms/step - accuracy: 0.4784 - loss: 1.1359
31/92 ━━━━━━━━━━━━━━━━━━━━ 15s 260ms/step - accuracy: 0.4787 - loss: 1.1350
32/92 ━━━━━━━━━━━━━━━━━━━━ 15s 260ms/step - accuracy: 0.4790 - loss: 1.1341
33/92 ━━━━━━━━━━━━━━━━━━━━ 15s 260ms/step - accuracy: 0.4794 - loss: 1.1334
34/92 ━━━━━━━━━━━━━━━━━━━━ 15s 260ms/step - accuracy: 0.4798 - loss: 1.1324
35/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4803 - loss: 1.1314
36/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4808 - loss: 1.1305
37/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4812 - loss: 1.1297
38/92 ━━━━━━━━━━━━━━━━━━━━ 14s 261ms/step - accuracy: 0.4816 - loss: 1.1289
39/92 ━━━━━━━━━━━━━━━━━━━━ 13s 261ms/step - accuracy: 0.4820 - loss: 1.1281
40/92 ━━━━━━━━━━━━━━━━━━━━ 13s 259ms/step - accuracy: 0.4824 - loss: 1.1273
41/92 ━━━━━━━━━━━━━━━━━━━━ 13s 259ms/step - accuracy: 0.4828 - loss: 1.1266
42/92 ━━━━━━━━━━━━━━━━━━━━ 12s 259ms/step - accuracy: 0.4831 - loss: 1.1259
43/92 ━━━━━━━━━━━━━━━━━━━━ 12s 259ms/step - accuracy: 0.4834 - loss: 1.1253
44/92 ━━━━━━━━━━━━━━━━━━━━ 12s 259ms/step - accuracy: 0.4837 - loss: 1.1246
45/92 ━━━━━━━━━━━━━━━━━━━━ 12s 259ms/step - accuracy: 0.4841 - loss: 1.1240
46/92 ━━━━━━━━━━━━━━━━━━━━ 11s 259ms/step - accuracy: 0.4844 - loss: 1.1233
47/92 ━━━━━━━━━━━━━━━━━━━━ 11s 258ms/step - accuracy: 0.4846 - loss: 1.1228
48/92 ━━━━━━━━━━━━━━━━━━━━ 11s 258ms/step - accuracy: 0.4849 - loss: 1.1223
49/92 ━━━━━━━━━━━━━━━━━━━━ 11s 258ms/step - accuracy: 0.4852 - loss: 1.1218
50/92 ━━━━━━━━━━━━━━━━━━━━ 10s 258ms/step - accuracy: 0.4855 - loss: 1.1211
51/92 ━━━━━━━━━━━━━━━━━━━━ 10s 258ms/step - accuracy: 0.4858 - loss: 1.1206
52/92 ━━━━━━━━━━━━━━━━━━━━ 10s 258ms/step - accuracy: 0.4861 - loss: 1.1201
53/92 ━━━━━━━━━━━━━━━━━━━━ 10s 259ms/step - accuracy: 0.4864 - loss: 1.1197
54/92 ━━━━━━━━━━━━━━━━━━━━ 9s 259ms/step - accuracy: 0.4866 - loss: 1.1192
55/92 ━━━━━━━━━━━━━━━━━━━━ 9s 259ms/step - accuracy: 0.4869 - loss: 1.1188
56/92 ━━━━━━━━━━━━━━━━━━━━ 9s 259ms/step - accuracy: 0.4871 - loss: 1.1183
57/92 ━━━━━━━━━━━━━━━━━━━━ 9s 259ms/step - accuracy: 0.4874 - loss: 1.1178
58/92 ━━━━━━━━━━━━━━━━━━━━ 8s 259ms/step - accuracy: 0.4876 - loss: 1.1174
59/92 ━━━━━━━━━━━━━━━━━━━━ 8s 259ms/step - accuracy: 0.4879 - loss: 1.1169
60/92 ━━━━━━━━━━━━━━━━━━━━ 8s 259ms/step - accuracy: 0.4882 - loss: 1.1164
61/92 ━━━━━━━━━━━━━━━━━━━━ 8s 259ms/step - accuracy: 0.4884 - loss: 1.1160
62/92 ━━━━━━━━━━━━━━━━━━━━ 7s 259ms/step - accuracy: 0.4886 - loss: 1.1157
63/92 ━━━━━━━━━━━━━━━━━━━━ 7s 259ms/step - accuracy: 0.4888 - loss: 1.1153
64/92 ━━━━━━━━━━━━━━━━━━━━ 7s 259ms/step - accuracy: 0.4891 - loss: 1.1149
65/92 ━━━━━━━━━━━━━━━━━━━━ 6s 259ms/step - accuracy: 0.4893 - loss: 1.1146
66/92 ━━━━━━━━━━━━━━━━━━━━ 6s 259ms/step - accuracy: 0.4895 - loss: 1.1142
67/92 ━━━━━━━━━━━━━━━━━━━━ 6s 259ms/step - accuracy: 0.4896 - loss: 1.1139
68/92 ━━━━━━━━━━━━━━━━━━━━ 6s 259ms/step - accuracy: 0.4898 - loss: 1.1136
69/92 ━━━━━━━━━━━━━━━━━━━━ 5s 259ms/step - accuracy: 0.4900 - loss: 1.1133
70/92 ━━━━━━━━━━━━━━━━━━━━ 5s 259ms/step - accuracy: 0.4901 - loss: 1.1129
71/92 ━━━━━━━━━━━━━━━━━━━━ 5s 259ms/step - accuracy: 0.4903 - loss: 1.1126
72/92 ━━━━━━━━━━━━━━━━━━━━ 5s 259ms/step - accuracy: 0.4904 - loss: 1.1122
73/92 ━━━━━━━━━━━━━━━━━━━━ 4s 259ms/step - accuracy: 0.4906 - loss: 1.1119
74/92 ━━━━━━━━━━━━━━━━━━━━ 4s 259ms/step - accuracy: 0.4907 - loss: 1.1115
75/92 ━━━━━━━━━━━━━━━━━━━━ 4s 259ms/step - accuracy: 0.4909 - loss: 1.1112
76/92 ━━━━━━━━━━━━━━━━━━━━ 4s 259ms/step - accuracy: 0.4910 - loss: 1.1109
77/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4912 - loss: 1.1106
78/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4913 - loss: 1.1103
79/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4914 - loss: 1.1100
80/92 ━━━━━━━━━━━━━━━━━━━━ 3s 259ms/step - accuracy: 0.4916 - loss: 1.1098
81/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4917 - loss: 1.1096
82/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4918 - loss: 1.1093
83/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4919 - loss: 1.1091
84/92 ━━━━━━━━━━━━━━━━━━━━ 2s 259ms/step - accuracy: 0.4920 - loss: 1.1089
85/92 ━━━━━━━━━━━━━━━━━━━━ 1s 259ms/step - accuracy: 0.4921 - loss: 1.1087
86/92 ━━━━━━━━━━━━━━━━━━━━ 1s 259ms/step - accuracy: 0.4922 - loss: 1.1085
87/92 ━━━━━━━━━━━━━━━━━━━━ 1s 260ms/step - accuracy: 0.4923 - loss: 1.1083
88/92 ━━━━━━━━━━━━━━━━━━━━ 1s 260ms/step - accuracy: 0.4924 - loss: 1.1081
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 260ms/step - accuracy: 0.4925 - loss: 1.1079
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 260ms/step - accuracy: 0.4926 - loss: 1.1078
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 261ms/step - accuracy: 0.4926 - loss: 1.1077
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 261ms/step - accuracy: 0.4927 - loss: 1.1075
92/92 ━━━━━━━━━━━━━━━━━━━━ 25s 274ms/step - accuracy: 0.5009 - loss: 1.0950 - val_accuracy: 0.3433 - val_loss: 1.1685
Epoch 13/15
 1/92 ━━━━━━━━━━━━━━━━━━━━ 25s 280ms/step - accuracy: 0.4844 - loss: 1.1733
 2/92 ━━━━━━━━━━━━━━━━━━━━ 25s 288ms/step - accuracy: 0.4727 - loss: 1.1606
 3/92 ━━━━━━━━━━━━━━━━━━━━ 24s 278ms/step - accuracy: 0.4696 - loss: 1.1531
 4/92 ━━━━━━━━━━━━━━━━━━━━ 24s 276ms/step - accuracy: 0.4665 - loss: 1.1451
 5/92 ━━━━━━━━━━━━━━━━━━━━ 23s 271ms/step - accuracy: 0.4744 - loss: 1.1270
 6/92 ━━━━━━━━━━━━━━━━━━━━ 23s 271ms/step - accuracy: 0.4783 - loss: 1.1181
 7/92 ━━━━━━━━━━━━━━━━━━━━ 22s 270ms/step - accuracy: 0.4817 - loss: 1.1138
 8/92 ━━━━━━━━━━━━━━━━━━━━ 22s 270ms/step - accuracy: 0.4852 - loss: 1.1095
 9/92 ━━━━━━━━━━━━━━━━━━━━ 22s 270ms/step - accuracy: 0.4897 - loss: 1.1026
10/92 ━━━━━━━━━━━━━━━━━━━━ 22s 271ms/step - accuracy: 0.4947 - loss: 1.0951
11/92 ━━━━━━━━━━━━━━━━━━━━ 21s 270ms/step - accuracy: 0.4986 - loss: 1.0883
12/92 ━━━━━━━━━━━━━━━━━━━━ 21s 270ms/step - accuracy: 0.5019 - loss: 1.0828
13/92 ━━━━━━━━━━━━━━━━━━━━ 21s 269ms/step - accuracy: 0.5042 - loss: 1.0785
14/92 ━━━━━━━━━━━━━━━━━━━━ 21s 271ms/step - accuracy: 0.5067 - loss: 1.0741
15/92 ━━━━━━━━━━━━━━━━━━━━ 20s 270ms/step - accuracy: 0.5090 - loss: 1.0698
16/92 ━━━━━━━━━━━━━━━━━━━━ 20s 269ms/step - accuracy: 0.5109 - loss: 1.0662
17/92 ━━━━━━━━━━━━━━━━━━━━ 20s 269ms/step - accuracy: 0.5126 - loss: 1.0633
18/92 ━━━━━━━━━━━━━━━━━━━━ 19s 268ms/step - accuracy: 0.5141 - loss: 1.0606
19/92 ━━━━━━━━━━━━━━━━━━━━ 19s 268ms/step - accuracy: 0.5153 - loss: 1.0584
20/92 ━━━━━━━━━━━━━━━━━━━━ 19s 268ms/step - accuracy: 0.5163 - loss: 1.0564
21/92 ━━━━━━━━━━━━━━━━━━━━ 19s 268ms/step - accuracy: 0.5170 - loss: 1.0549
22/92 ━━━━━━━━━━━━━━━━━━━━ 18s 268ms/step - accuracy: 0.5177 - loss: 1.0533
23/92 ━━━━━━━━━━━━━━━━━━━━ 18s 268ms/step - accuracy: 0.5181 - loss: 1.0522
24/92 ━━━━━━━━━━━━━━━━━━━━ 18s 268ms/step - accuracy: 0.5185 - loss: 1.0511
25/92 ━━━━━━━━━━━━━━━━━━━━ 17s 268ms/step - accuracy: 0.5190 - loss: 1.0498
26/92 ━━━━━━━━━━━━━━━━━━━━ 17s 268ms/step - accuracy: 0.5195 - loss: 1.0486
27/92 ━━━━━━━━━━━━━━━━━━━━ 17s 268ms/step - accuracy: 0.5199 - loss: 1.0479
28/92 ━━━━━━━━━━━━━━━━━━━━ 17s 268ms/step - accuracy: 0.5203 - loss: 1.0471
29/92 ━━━━━━━━━━━━━━━━━━━━ 16s 268ms/step - accuracy: 0.5206 - loss: 1.0463
30/92 ━━━━━━━━━━━━━━━━━━━━ 16s 268ms/step - accuracy: 0.5210 - loss: 1.0455
31/92 ━━━━━━━━━━━━━━━━━━━━ 16s 268ms/step - accuracy: 0.5215 - loss: 1.0446
32/92 ━━━━━━━━━━━━━━━━━━━━ 16s 268ms/step - accuracy: 0.5217 - loss: 1.0439
33/92 ━━━━━━━━━━━━━━━━━━━━ 15s 268ms/step - accuracy: 0.5220 - loss: 1.0433
34/92 ━━━━━━━━━━━━━━━━━━━━ 15s 269ms/step - accuracy: 0.5222 - loss: 1.0427
35/92 ━━━━━━━━━━━━━━━━━━━━ 15s 269ms/step - accuracy: 0.5224 - loss: 1.0422
36/92 ━━━━━━━━━━━━━━━━━━━━ 15s 269ms/step - accuracy: 0.5226 - loss: 1.0417
37/92 ━━━━━━━━━━━━━━━━━━━━ 14s 269ms/step - accuracy: 0.5227 - loss: 1.0411
38/92 ━━━━━━━━━━━━━━━━━━━━ 14s 268ms/step - accuracy: 0.5229 - loss: 1.0405
39/92 ━━━━━━━━━━━━━━━━━━━━ 14s 268ms/step - accuracy: 0.5230 - loss: 1.0400
40/92 ━━━━━━━━━━━━━━━━━━━━ 13s 269ms/step - accuracy: 0.5232 - loss: 1.0395
41/92 ━━━━━━━━━━━━━━━━━━━━ 13s 268ms/step - accuracy: 0.5234 - loss: 1.0391
42/92 ━━━━━━━━━━━━━━━━━━━━ 13s 268ms/step - accuracy: 0.5235 - loss: 1.0387
43/92 ━━━━━━━━━━━━━━━━━━━━ 13s 268ms/step - accuracy: 0.5237 - loss: 1.0383
44/92 ━━━━━━━━━━━━━━━━━━━━ 12s 268ms/step - accuracy: 0.5238 - loss: 1.0379
45/92 ━━━━━━━━━━━━━━━━━━━━ 12s 268ms/step - accuracy: 0.5239 - loss: 1.0376
46/92 ━━━━━━━━━━━━━━━━━━━━ 12s 268ms/step - accuracy: 0.5241 - loss: 1.0372
47/92 ━━━━━━━━━━━━━━━━━━━━ 12s 267ms/step - accuracy: 0.5242 - loss: 1.0368
48/92 ━━━━━━━━━━━━━━━━━━━━ 11s 267ms/step - accuracy: 0.5244 - loss: 1.0365
49/92 ━━━━━━━━━━━━━━━━━━━━ 11s 267ms/step - accuracy: 0.5246 - loss: 1.0361
50/92 ━━━━━━━━━━━━━━━━━━━━ 11s 267ms/step - accuracy: 0.5247 - loss: 1.0358
51/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5249 - loss: 1.0355
52/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5250 - loss: 1.0351
53/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5252 - loss: 1.0348
54/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5253 - loss: 1.0345
55/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5254 - loss: 1.0341
56/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5256 - loss: 1.0338
57/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5257 - loss: 1.0335
58/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5258 - loss: 1.0331
59/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5259 - loss: 1.0329
60/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5260 - loss: 1.0326
61/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5261 - loss: 1.0323
62/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5262 - loss: 1.0321
63/92 ━━━━━━━━━━━━━━━━━━━━ 7s 267ms/step - accuracy: 0.5263 - loss: 1.0318
64/92 ━━━━━━━━━━━━━━━━━━━━ 7s 267ms/step - accuracy: 0.5264 - loss: 1.0315
65/92 ━━━━━━━━━━━━━━━━━━━━ 7s 267ms/step - accuracy: 0.5265 - loss: 1.0312
66/92 ━━━━━━━━━━━━━━━━━━━━ 6s 267ms/step - accuracy: 0.5266 - loss: 1.0310
67/92 ━━━━━━━━━━━━━━━━━━━━ 6s 267ms/step - accuracy: 0.5266 - loss: 1.0307
68/92 ━━━━━━━━━━━━━━━━━━━━ 6s 267ms/step - accuracy: 0.5267 - loss: 1.0305
69/92 ━━━━━━━━━━━━━━━━━━━━ 6s 266ms/step - accuracy: 0.5268 - loss: 1.0303
70/92 ━━━━━━━━━━━━━━━━━━━━ 5s 266ms/step - accuracy: 0.5269 - loss: 1.0301
71/92 ━━━━━━━━━━━━━━━━━━━━ 5s 266ms/step - accuracy: 0.5270 - loss: 1.0299
72/92 ━━━━━━━━━━━━━━━━━━━━ 5s 266ms/step - accuracy: 0.5271 - loss: 1.0296
73/92 ━━━━━━━━━━━━━━━━━━━━ 5s 266ms/step - accuracy: 0.5271 - loss: 1.0295
74/92 ━━━━━━━━━━━━━━━━━━━━ 4s 267ms/step - accuracy: 0.5272 - loss: 1.0293
75/92 ━━━━━━━━━━━━━━━━━━━━ 4s 267ms/step - accuracy: 0.5273 - loss: 1.0292
76/92 ━━━━━━━━━━━━━━━━━━━━ 4s 266ms/step - accuracy: 0.5274 - loss: 1.0291
77/92 ━━━━━━━━━━━━━━━━━━━━ 3s 266ms/step - accuracy: 0.5274 - loss: 1.0289
78/92 ━━━━━━━━━━━━━━━━━━━━ 3s 266ms/step - accuracy: 0.5275 - loss: 1.0288
79/92 ━━━━━━━━━━━━━━━━━━━━ 3s 266ms/step - accuracy: 0.5276 - loss: 1.0287
80/92 ━━━━━━━━━━━━━━━━━━━━ 3s 266ms/step - accuracy: 0.5276 - loss: 1.0286
81/92 ━━━━━━━━━━━━━━━━━━━━ 2s 266ms/step - accuracy: 0.5277 - loss: 1.0285
82/92 ━━━━━━━━━━━━━━━━━━━━ 2s 266ms/step - accuracy: 0.5277 - loss: 1.0284
83/92 ━━━━━━━━━━━━━━━━━━━━ 2s 266ms/step - accuracy: 0.5278 - loss: 1.0283
84/92 ━━━━━━━━━━━━━━━━━━━━ 2s 266ms/step - accuracy: 0.5278 - loss: 1.0282
85/92 ━━━━━━━━━━━━━━━━━━━━ 1s 266ms/step - accuracy: 0.5279 - loss: 1.0281
86/92 ━━━━━━━━━━━━━━━━━━━━ 1s 266ms/step - accuracy: 0.5279 - loss: 1.0280
87/92 ━━━━━━━━━━━━━━━━━━━━ 1s 266ms/step - accuracy: 0.5280 - loss: 1.0279
88/92 ━━━━━━━━━━━━━━━━━━━━ 1s 266ms/step - accuracy: 0.5280 - loss: 1.0279
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 266ms/step - accuracy: 0.5280 - loss: 1.0278
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 266ms/step - accuracy: 0.5280 - loss: 1.0277
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 266ms/step - accuracy: 0.5280 - loss: 1.0277
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 266ms/step - accuracy: 0.5280 - loss: 1.0276
92/92 ━━━━━━━━━━━━━━━━━━━━ 26s 279ms/step - accuracy: 0.5278 - loss: 1.0226 - val_accuracy: 0.3481 - val_loss: 1.1748
Epoch 14/15
 1/92 ━━━━━━━━━━━━━━━━━━━━ 24s 271ms/step - accuracy: 0.5781 - loss: 0.9226
 2/92 ━━━━━━━━━━━━━━━━━━━━ 23s 265ms/step - accuracy: 0.5664 - loss: 0.9334
 3/92 ━━━━━━━━━━━━━━━━━━━━ 22s 258ms/step - accuracy: 0.5703 - loss: 0.9297
 4/92 ━━━━━━━━━━━━━━━━━━━━ 22s 260ms/step - accuracy: 0.5723 - loss: 0.9258
 5/92 ━━━━━━━━━━━━━━━━━━━━ 22s 261ms/step - accuracy: 0.5697 - loss: 0.9241
 6/92 ━━━━━━━━━━━━━━━━━━━━ 22s 263ms/step - accuracy: 0.5668 - loss: 0.9249
 7/92 ━━━━━━━━━━━━━━━━━━━━ 22s 263ms/step - accuracy: 0.5649 - loss: 0.9278
 8/92 ━━━━━━━━━━━━━━━━━━━━ 22s 265ms/step - accuracy: 0.5636 - loss: 0.9312
 9/92 ━━━━━━━━━━━━━━━━━━━━ 21s 264ms/step - accuracy: 0.5619 - loss: 0.9340
10/92 ━━━━━━━━━━━━━━━━━━━━ 21s 264ms/step - accuracy: 0.5607 - loss: 0.9359
11/92 ━━━━━━━━━━━━━━━━━━━━ 21s 263ms/step - accuracy: 0.5588 - loss: 0.9379
12/92 ━━━━━━━━━━━━━━━━━━━━ 21s 263ms/step - accuracy: 0.5573 - loss: 0.9390
13/92 ━━━━━━━━━━━━━━━━━━━━ 20s 263ms/step - accuracy: 0.5557 - loss: 0.9408
14/92 ━━━━━━━━━━━━━━━━━━━━ 20s 263ms/step - accuracy: 0.5541 - loss: 0.9427
15/92 ━━━━━━━━━━━━━━━━━━━━ 20s 263ms/step - accuracy: 0.5527 - loss: 0.9445
16/92 ━━━━━━━━━━━━━━━━━━━━ 19s 262ms/step - accuracy: 0.5512 - loss: 0.9461
17/92 ━━━━━━━━━━━━━━━━━━━━ 19s 263ms/step - accuracy: 0.5499 - loss: 0.9472
18/92 ━━━━━━━━━━━━━━━━━━━━ 19s 263ms/step - accuracy: 0.5489 - loss: 0.9480
19/92 ━━━━━━━━━━━━━━━━━━━━ 19s 262ms/step - accuracy: 0.5483 - loss: 0.9488
20/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.5476 - loss: 0.9494
21/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.5469 - loss: 0.9503
22/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.5464 - loss: 0.9511
23/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.5460 - loss: 0.9518
24/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.5455 - loss: 0.9525
25/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.5451 - loss: 0.9530
26/92 ━━━━━━━━━━━━━━━━━━━━ 17s 263ms/step - accuracy: 0.5448 - loss: 0.9535
27/92 ━━━━━━━━━━━━━━━━━━━━ 17s 263ms/step - accuracy: 0.5445 - loss: 0.9540
28/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.5442 - loss: 0.9544
29/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.5439 - loss: 0.9548
30/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.5436 - loss: 0.9552
31/92 ━━━━━━━━━━━━━━━━━━━━ 15s 262ms/step - accuracy: 0.5432 - loss: 0.9556
32/92 ━━━━━━━━━━━━━━━━━━━━ 15s 263ms/step - accuracy: 0.5429 - loss: 0.9557
33/92 ━━━━━━━━━━━━━━━━━━━━ 15s 262ms/step - accuracy: 0.5427 - loss: 0.9559
34/92 ━━━━━━━━━━━━━━━━━━━━ 15s 263ms/step - accuracy: 0.5424 - loss: 0.9562
35/92 ━━━━━━━━━━━━━━━━━━━━ 14s 263ms/step - accuracy: 0.5421 - loss: 0.9565
36/92 ━━━━━━━━━━━━━━━━━━━━ 14s 263ms/step - accuracy: 0.5420 - loss: 0.9567
37/92 ━━━━━━━━━━━━━━━━━━━━ 14s 263ms/step - accuracy: 0.5419 - loss: 0.9569
38/92 ━━━━━━━━━━━━━━━━━━━━ 14s 263ms/step - accuracy: 0.5419 - loss: 0.9571
39/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5418 - loss: 0.9572
40/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5417 - loss: 0.9574
41/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5417 - loss: 0.9576
42/92 ━━━━━━━━━━━━━━━━━━━━ 13s 262ms/step - accuracy: 0.5415 - loss: 0.9578
43/92 ━━━━━━━━━━━━━━━━━━━━ 12s 262ms/step - accuracy: 0.5413 - loss: 0.9580
44/92 ━━━━━━━━━━━━━━━━━━━━ 12s 262ms/step - accuracy: 0.5413 - loss: 0.9582
45/92 ━━━━━━━━━━━━━━━━━━━━ 12s 262ms/step - accuracy: 0.5412 - loss: 0.9582
46/92 ━━━━━━━━━━━━━━━━━━━━ 12s 262ms/step - accuracy: 0.5411 - loss: 0.9583
47/92 ━━━━━━━━━━━━━━━━━━━━ 11s 262ms/step - accuracy: 0.5410 - loss: 0.9583
48/92 ━━━━━━━━━━━━━━━━━━━━ 11s 262ms/step - accuracy: 0.5410 - loss: 0.9584
49/92 ━━━━━━━━━━━━━━━━━━━━ 11s 262ms/step - accuracy: 0.5409 - loss: 0.9585
50/92 ━━━━━━━━━━━━━━━━━━━━ 11s 262ms/step - accuracy: 0.5409 - loss: 0.9584
51/92 ━━━━━━━━━━━━━━━━━━━━ 10s 262ms/step - accuracy: 0.5410 - loss: 0.9583
52/92 ━━━━━━━━━━━━━━━━━━━━ 10s 262ms/step - accuracy: 0.5410 - loss: 0.9582
53/92 ━━━━━━━━━━━━━━━━━━━━ 10s 262ms/step - accuracy: 0.5410 - loss: 0.9581
54/92 ━━━━━━━━━━━━━━━━━━━━ 9s 262ms/step - accuracy: 0.5410 - loss: 0.9581
55/92 ━━━━━━━━━━━━━━━━━━━━ 9s 262ms/step - accuracy: 0.5410 - loss: 0.9581
56/92 ━━━━━━━━━━━━━━━━━━━━ 9s 262ms/step - accuracy: 0.5410 - loss: 0.9580
57/92 ━━━━━━━━━━━━━━━━━━━━ 9s 262ms/step - accuracy: 0.5410 - loss: 0.9580
58/92 ━━━━━━━━━━━━━━━━━━━━ 8s 262ms/step - accuracy: 0.5410 - loss: 0.9580
59/92 ━━━━━━━━━━━━━━━━━━━━ 8s 262ms/step - accuracy: 0.5410 - loss: 0.9580
60/92 ━━━━━━━━━━━━━━━━━━━━ 8s 262ms/step - accuracy: 0.5410 - loss: 0.9579
61/92 ━━━━━━━━━━━━━━━━━━━━ 8s 262ms/step - accuracy: 0.5410 - loss: 0.9580
62/92 ━━━━━━━━━━━━━━━━━━━━ 7s 262ms/step - accuracy: 0.5410 - loss: 0.9579
63/92 ━━━━━━━━━━━━━━━━━━━━ 7s 262ms/step - accuracy: 0.5410 - loss: 0.9579
64/92 ━━━━━━━━━━━━━━━━━━━━ 7s 262ms/step - accuracy: 0.5410 - loss: 0.9579
65/92 ━━━━━━━━━━━━━━━━━━━━ 7s 262ms/step - accuracy: 0.5410 - loss: 0.9579
66/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.5411 - loss: 0.9578
67/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.5411 - loss: 0.9577
68/92 ━━━━━━━━━━━━━━━━━━━━ 6s 262ms/step - accuracy: 0.5411 - loss: 0.9576
69/92 ━━━━━━━━━━━━━━━━━━━━ 6s 261ms/step - accuracy: 0.5412 - loss: 0.9575
70/92 ━━━━━━━━━━━━━━━━━━━━ 5s 261ms/step - accuracy: 0.5412 - loss: 0.9574
71/92 ━━━━━━━━━━━━━━━━━━━━ 5s 261ms/step - accuracy: 0.5413 - loss: 0.9572
72/92 ━━━━━━━━━━━━━━━━━━━━ 5s 261ms/step - accuracy: 0.5414 - loss: 0.9572
73/92 ━━━━━━━━━━━━━━━━━━━━ 4s 261ms/step - accuracy: 0.5414 - loss: 0.9571
74/92 ━━━━━━━━━━━━━━━━━━━━ 4s 261ms/step - accuracy: 0.5414 - loss: 0.9570
75/92 ━━━━━━━━━━━━━━━━━━━━ 4s 261ms/step - accuracy: 0.5415 - loss: 0.9569
76/92 ━━━━━━━━━━━━━━━━━━━━ 4s 261ms/step - accuracy: 0.5415 - loss: 0.9569
77/92 ━━━━━━━━━━━━━━━━━━━━ 3s 261ms/step - accuracy: 0.5415 - loss: 0.9568
78/92 ━━━━━━━━━━━━━━━━━━━━ 3s 261ms/step - accuracy: 0.5415 - loss: 0.9568
79/92 ━━━━━━━━━━━━━━━━━━━━ 3s 261ms/step - accuracy: 0.5415 - loss: 0.9568
80/92 ━━━━━━━━━━━━━━━━━━━━ 3s 261ms/step - accuracy: 0.5415 - loss: 0.9567
81/92 ━━━━━━━━━━━━━━━━━━━━ 2s 261ms/step - accuracy: 0.5415 - loss: 0.9567
82/92 ━━━━━━━━━━━━━━━━━━━━ 2s 261ms/step - accuracy: 0.5415 - loss: 0.9566
83/92 ━━━━━━━━━━━━━━━━━━━━ 2s 261ms/step - accuracy: 0.5415 - loss: 0.9566
84/92 ━━━━━━━━━━━━━━━━━━━━ 2s 262ms/step - accuracy: 0.5415 - loss: 0.9565
85/92 ━━━━━━━━━━━━━━━━━━━━ 1s 262ms/step - accuracy: 0.5415 - loss: 0.9565
86/92 ━━━━━━━━━━━━━━━━━━━━ 1s 262ms/step - accuracy: 0.5415 - loss: 0.9564
87/92 ━━━━━━━━━━━━━━━━━━━━ 1s 262ms/step - accuracy: 0.5415 - loss: 0.9563
88/92 ━━━━━━━━━━━━━━━━━━━━ 1s 262ms/step - accuracy: 0.5416 - loss: 0.9563
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 262ms/step - accuracy: 0.5416 - loss: 0.9562
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 262ms/step - accuracy: 0.5416 - loss: 0.9561
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 262ms/step - accuracy: 0.5416 - loss: 0.9561
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 262ms/step - accuracy: 0.5415 - loss: 0.9561
92/92 ━━━━━━━━━━━━━━━━━━━━ 25s 274ms/step - accuracy: 0.5395 - loss: 0.9571 - val_accuracy: 0.3399 - val_loss: 1.1813
Epoch 15/15
 1/92 ━━━━━━━━━━━━━━━━━━━━ 23s 261ms/step - accuracy: 0.5312 - loss: 0.9057
 2/92 ━━━━━━━━━━━━━━━━━━━━ 23s 258ms/step - accuracy: 0.5430 - loss: 0.9036
 3/92 ━━━━━━━━━━━━━━━━━━━━ 23s 261ms/step - accuracy: 0.5425 - loss: 0.9202
 4/92 ━━━━━━━━━━━━━━━━━━━━ 22s 260ms/step - accuracy: 0.5426 - loss: 0.9249
 5/92 ━━━━━━━━━━━━━━━━━━━━ 21s 245ms/step - accuracy: 0.5446 - loss: 0.9250
 6/92 ━━━━━━━━━━━━━━━━━━━━ 21s 250ms/step - accuracy: 0.5458 - loss: 0.9264
 7/92 ━━━━━━━━━━━━━━━━━━━━ 21s 252ms/step - accuracy: 0.5469 - loss: 0.9261
 8/92 ━━━━━━━━━━━━━━━━━━━━ 21s 253ms/step - accuracy: 0.5466 - loss: 0.9267
 9/92 ━━━━━━━━━━━━━━━━━━━━ 21s 254ms/step - accuracy: 0.5473 - loss: 0.9264
10/92 ━━━━━━━━━━━━━━━━━━━━ 20s 254ms/step - accuracy: 0.5474 - loss: 0.9265
11/92 ━━━━━━━━━━━━━━━━━━━━ 21s 261ms/step - accuracy: 0.5479 - loss: 0.9269
12/92 ━━━━━━━━━━━━━━━━━━━━ 20s 262ms/step - accuracy: 0.5487 - loss: 0.9262
13/92 ━━━━━━━━━━━━━━━━━━━━ 20s 263ms/step - accuracy: 0.5502 - loss: 0.9248
14/92 ━━━━━━━━━━━━━━━━━━━━ 20s 264ms/step - accuracy: 0.5515 - loss: 0.9235
15/92 ━━━━━━━━━━━━━━━━━━━━ 20s 264ms/step - accuracy: 0.5525 - loss: 0.9227
16/92 ━━━━━━━━━━━━━━━━━━━━ 20s 264ms/step - accuracy: 0.5529 - loss: 0.9223
17/92 ━━━━━━━━━━━━━━━━━━━━ 19s 263ms/step - accuracy: 0.5532 - loss: 0.9220
18/92 ━━━━━━━━━━━━━━━━━━━━ 19s 263ms/step - accuracy: 0.5532 - loss: 0.9220
19/92 ━━━━━━━━━━━━━━━━━━━━ 19s 263ms/step - accuracy: 0.5535 - loss: 0.9217
20/92 ━━━━━━━━━━━━━━━━━━━━ 18s 263ms/step - accuracy: 0.5537 - loss: 0.9213
21/92 ━━━━━━━━━━━━━━━━━━━━ 18s 263ms/step - accuracy: 0.5540 - loss: 0.9209
22/92 ━━━━━━━━━━━━━━━━━━━━ 18s 263ms/step - accuracy: 0.5541 - loss: 0.9208
23/92 ━━━━━━━━━━━━━━━━━━━━ 18s 262ms/step - accuracy: 0.5541 - loss: 0.9207
24/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.5543 - loss: 0.9205
25/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.5543 - loss: 0.9202
26/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.5545 - loss: 0.9198
27/92 ━━━━━━━━━━━━━━━━━━━━ 17s 262ms/step - accuracy: 0.5547 - loss: 0.9194
28/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.5549 - loss: 0.9189
29/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.5551 - loss: 0.9186
30/92 ━━━━━━━━━━━━━━━━━━━━ 16s 262ms/step - accuracy: 0.5552 - loss: 0.9184
31/92 ━━━━━━━━━━━━━━━━━━━━ 16s 263ms/step - accuracy: 0.5553 - loss: 0.9182
32/92 ━━━━━━━━━━━━━━━━━━━━ 15s 263ms/step - accuracy: 0.5554 - loss: 0.9180
33/92 ━━━━━━━━━━━━━━━━━━━━ 15s 263ms/step - accuracy: 0.5555 - loss: 0.9180
34/92 ━━━━━━━━━━━━━━━━━━━━ 15s 263ms/step - accuracy: 0.5556 - loss: 0.9180
35/92 ━━━━━━━━━━━━━━━━━━━━ 15s 263ms/step - accuracy: 0.5556 - loss: 0.9180
36/92 ━━━━━━━━━━━━━━━━━━━━ 14s 264ms/step - accuracy: 0.5557 - loss: 0.9180
37/92 ━━━━━━━━━━━━━━━━━━━━ 14s 263ms/step - accuracy: 0.5557 - loss: 0.9180
38/92 ━━━━━━━━━━━━━━━━━━━━ 14s 263ms/step - accuracy: 0.5557 - loss: 0.9182
39/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5557 - loss: 0.9184
40/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5557 - loss: 0.9186
41/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5557 - loss: 0.9187
42/92 ━━━━━━━━━━━━━━━━━━━━ 13s 263ms/step - accuracy: 0.5557 - loss: 0.9189
43/92 ━━━━━━━━━━━━━━━━━━━━ 12s 263ms/step - accuracy: 0.5556 - loss: 0.9191
44/92 ━━━━━━━━━━━━━━━━━━━━ 12s 263ms/step - accuracy: 0.5557 - loss: 0.9193
45/92 ━━━━━━━━━━━━━━━━━━━━ 12s 263ms/step - accuracy: 0.5557 - loss: 0.9194
46/92 ━━━━━━━━━━━━━━━━━━━━ 12s 263ms/step - accuracy: 0.5557 - loss: 0.9195
47/92 ━━━━━━━━━━━━━━━━━━━━ 11s 263ms/step - accuracy: 0.5558 - loss: 0.9195
48/92 ━━━━━━━━━━━━━━━━━━━━ 11s 264ms/step - accuracy: 0.5558 - loss: 0.9196
49/92 ━━━━━━━━━━━━━━━━━━━━ 11s 265ms/step - accuracy: 0.5558 - loss: 0.9197
50/92 ━━━━━━━━━━━━━━━━━━━━ 11s 266ms/step - accuracy: 0.5558 - loss: 0.9198
51/92 ━━━━━━━━━━━━━━━━━━━━ 10s 266ms/step - accuracy: 0.5558 - loss: 0.9198
52/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5559 - loss: 0.9199
53/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5559 - loss: 0.9199
54/92 ━━━━━━━━━━━━━━━━━━━━ 10s 267ms/step - accuracy: 0.5559 - loss: 0.9199
55/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5560 - loss: 0.9199
56/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5560 - loss: 0.9198
57/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5561 - loss: 0.9198
58/92 ━━━━━━━━━━━━━━━━━━━━ 9s 267ms/step - accuracy: 0.5561 - loss: 0.9197
59/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5562 - loss: 0.9196
60/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5563 - loss: 0.9195
61/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5564 - loss: 0.9193
62/92 ━━━━━━━━━━━━━━━━━━━━ 8s 267ms/step - accuracy: 0.5565 - loss: 0.9192
63/92 ━━━━━━━━━━━━━━━━━━━━ 7s 267ms/step - accuracy: 0.5566 - loss: 0.9191
64/92 ━━━━━━━━━━━━━━━━━━━━ 7s 268ms/step - accuracy: 0.5566 - loss: 0.9190
65/92 ━━━━━━━━━━━━━━━━━━━━ 7s 268ms/step - accuracy: 0.5566 - loss: 0.9190
66/92 ━━━━━━━━━━━━━━━━━━━━ 6s 268ms/step - accuracy: 0.5566 - loss: 0.9190
67/92 ━━━━━━━━━━━━━━━━━━━━ 6s 268ms/step - accuracy: 0.5566 - loss: 0.9189
68/92 ━━━━━━━━━━━━━━━━━━━━ 6s 269ms/step - accuracy: 0.5566 - loss: 0.9189
69/92 ━━━━━━━━━━━━━━━━━━━━ 6s 269ms/step - accuracy: 0.5566 - loss: 0.9188
70/92 ━━━━━━━━━━━━━━━━━━━━ 5s 270ms/step - accuracy: 0.5567 - loss: 0.9188
71/92 ━━━━━━━━━━━━━━━━━━━━ 5s 271ms/step - accuracy: 0.5567 - loss: 0.9187
72/92 ━━━━━━━━━━━━━━━━━━━━ 5s 273ms/step - accuracy: 0.5567 - loss: 0.9187
73/92 ━━━━━━━━━━━━━━━━━━━━ 5s 274ms/step - accuracy: 0.5567 - loss: 0.9186
74/92 ━━━━━━━━━━━━━━━━━━━━ 4s 274ms/step - accuracy: 0.5567 - loss: 0.9186
75/92 ━━━━━━━━━━━━━━━━━━━━ 4s 274ms/step - accuracy: 0.5567 - loss: 0.9186
76/92 ━━━━━━━━━━━━━━━━━━━━ 4s 274ms/step - accuracy: 0.5567 - loss: 0.9186
77/92 ━━━━━━━━━━━━━━━━━━━━ 4s 275ms/step - accuracy: 0.5568 - loss: 0.9185
78/92 ━━━━━━━━━━━━━━━━━━━━ 3s 275ms/step - accuracy: 0.5568 - loss: 0.9185
79/92 ━━━━━━━━━━━━━━━━━━━━ 3s 275ms/step - accuracy: 0.5568 - loss: 0.9184
80/92 ━━━━━━━━━━━━━━━━━━━━ 3s 275ms/step - accuracy: 0.5568 - loss: 0.9184
81/92 ━━━━━━━━━━━━━━━━━━━━ 3s 275ms/step - accuracy: 0.5568 - loss: 0.9183
82/92 ━━━━━━━━━━━━━━━━━━━━ 2s 275ms/step - accuracy: 0.5568 - loss: 0.9183
83/92 ━━━━━━━━━━━━━━━━━━━━ 2s 275ms/step - accuracy: 0.5568 - loss: 0.9183
84/92 ━━━━━━━━━━━━━━━━━━━━ 2s 275ms/step - accuracy: 0.5568 - loss: 0.9182
85/92 ━━━━━━━━━━━━━━━━━━━━ 1s 274ms/step - accuracy: 0.5568 - loss: 0.9182
86/92 ━━━━━━━━━━━━━━━━━━━━ 1s 274ms/step - accuracy: 0.5568 - loss: 0.9182
87/92 ━━━━━━━━━━━━━━━━━━━━ 1s 274ms/step - accuracy: 0.5568 - loss: 0.9182
88/92 ━━━━━━━━━━━━━━━━━━━━ 1s 274ms/step - accuracy: 0.5568 - loss: 0.9182
89/92 ━━━━━━━━━━━━━━━━━━━━ 0s 274ms/step - accuracy: 0.5567 - loss: 0.9182
90/92 ━━━━━━━━━━━━━━━━━━━━ 0s 274ms/step - accuracy: 0.5567 - loss: 0.9183
91/92 ━━━━━━━━━━━━━━━━━━━━ 0s 274ms/step - accuracy: 0.5567 - loss: 0.9183
92/92 ━━━━━━━━━━━━━━━━━━━━ 0s 274ms/step - accuracy: 0.5566 - loss: 0.9183
92/92 ━━━━━━━━━━━━━━━━━━━━ 26s 287ms/step - accuracy: 0.5536 - loss: 0.9181 - val_accuracy: 0.3358 - val_loss: 1.1926

Visualizing Fine-tuning Results

plot_training_history(history_finetune, 'Fine-tuning')

val_loss_finetune, val_acc_finetune = model_transfer.evaluate(val_ds, verbose=0)
print(f'\nFine-tuning Results:')
print(f'  Validation Loss: {val_loss_finetune:.4f}')
print(f'  Validation Accuracy: {val_acc_finetune:.4f}')

Fine-tuning Results:
  Validation Loss: 1.1926
  Validation Accuracy: 0.3358

Combined Training Curve

Let’s visualize the complete training process: frozen backbone + fine-tuning.

# Combine histories from both training phases
acc = history_transfer.history['accuracy'] + history_finetune.history['accuracy']
val_acc = history_transfer.history['val_accuracy'] + history_finetune.history['val_accuracy']
loss = history_transfer.history['loss'] + history_finetune.history['loss']
val_loss = history_transfer.history['val_loss'] + history_finetune.history['val_loss']

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))

# Loss plot
ax1.plot(loss, label='Train Loss', marker='o')
ax1.plot(val_loss, label='Val Loss', marker='s')
ax1.axvline(x=epochs_initial-1, color='r', linestyle='--', label='Start Fine-tuning')  # Mark transition
ax1.set_xlabel('Epoch')
ax1.set_ylabel('Loss')
ax1.set_title('Complete Training: Frozen + Fine-tuning')
ax1.legend()
ax1.grid(True, alpha=0.3)

# Accuracy plot
ax2.plot(acc, label='Train Accuracy', marker='o')
ax2.plot(val_acc, label='Val Accuracy', marker='s')
ax2.axvline(x=epochs_initial-1, color='r', linestyle='--', label='Start Fine-tuning')  # Mark transition
ax2.set_xlabel('Epoch')
ax2.set_ylabel('Accuracy')
ax2.set_title('Complete Training: Frozen + Fine-tuning')
ax2.legend()
ax2.grid(True, alpha=0.3)

plt.tight_layout()
plt.show()

Comparing Approaches

Let’s summarize the different approaches we could use:

Training from Scratch (Baseline CNN)

• Train all weights randomly initialized

• Requires more data and epochs

• Higher computational cost

Transfer Learning (Frozen)

• Freeze pre-trained backbone

• Train only new classification layers

• Fast training, good baseline results

Hybrid Model

• Combines custom CNN with pre-trained features

• Learns task-specific features while leveraging ImageNet knowledge

• More parameters than frozen, less than full fine-tuning

Fine-tuning

• Start with frozen backbone training

• Then unfreeze and train with low learning rate

• Best accuracy, moderate training time

Comparison Table

Approach	Trainable Params	Training Time	Typical Accuracy	When to Use
From scratch (baseline)	All (~500K-2M)	Slow	Baseline	Large dataset, unique domain
Frozen backbone	Few (~10K-50K)	Fast	Good	Limited data, similar to ImageNet
Hybrid model	Medium (~50K-200K)	Medium	Good-Better	Dataset with unique characteristics
Fine-tuning	All (~500K-3M)	Medium	Best	Want highest accuracy, have some data

Performance Summary

print("Performance Comparison:")
print("="*60)
print(f"Transfer Learning (Frozen):  {val_acc_transfer:.4f}")
print(f"Hybrid Model:                {val_acc_hybrid:.4f}")
print(f"Fine-tuning:                 {val_acc_finetune:.4f}")
print("="*60)
print(f"\nImprovement from fine-tuning: {(val_acc_finetune - val_acc_transfer):.4f}")

Performance Comparison:
============================================================
Transfer Learning (Frozen):  0.3529
Hybrid Model:                0.3297
Fine-tuning:                 0.3358
============================================================

Improvement from fine-tuning: -0.0170

Typical observations:

• Transfer learning achieves good results quickly (70-80% accuracy in 10 epochs)

• Hybrid models can match or exceed frozen transfer learning

• Fine-tuning usually improves accuracy by 2-5%

• All approaches vastly outperform training from scratch with limited data

Exercises

Exercise 1: Freeze/Unfreeze Experiment

Compare training with backbone frozen vs unfrozen.

Tasks:

1. Train with base_model.trainable = False (feature extraction)

2. Train with base_model.trainable = True (fine-tuning)

3. Compare validation accuracy and training time

Key API:

• base_model.trainable = False freezes all weights

• layer.trainable = False freezes individual layers

# Exercise 1: Compare frozen vs unfrozen backbone
# TODO: Set base_model.trainable = False, train for 5 epochs, note val_accuracy
# TODO: Then set base_model.trainable = True, train for 5 epochs, compare
# 
# Hint: With unfrozen backbone, use a lower learning rate (1e-5 vs 1e-3)

Exercise 2: Learning Rate for Fine-tuning

Experiment with learning rates when fine-tuning.

Tasks:

1. Fine-tune with lr=1e-3 (high) - what happens?

2. Fine-tune with lr=1e-5 (low) - compare results

Why it matters:

• High LR can destroy pre-trained features (“catastrophic forgetting”)

• Low LR preserves features but updates slowly

Key API:

• model.compile(optimizer=Adam(learning_rate=...))

# Exercise 2: Learning rate comparison
# TODO: Train with high LR (1e-3) and low LR (1e-5) with unfrozen backbone
# Watch for signs of catastrophic forgetting (val accuracy drops suddenly)
# 
# from tensorflow.keras.optimizers import Adam
# model.compile(optimizer=Adam(learning_rate=1e-5), ...)

Recap

Transfer Learning Concept:

• Reuse knowledge from models trained on large datasets (ImageNet with 1.4M images)

• Reduces data requirements and training time significantly

• Often achieves better results than training from scratch, especially with limited data

Pre-trained Models:

• Keras Applications provides many CNN architectures (MobileNetV2, EfficientNet, ResNet)

• Models pre-trained on ImageNet (1000 classes: animals, plants, objects)

• Different models offer tradeoffs: MobileNetV2 (fast), ResNet50 (accurate), EfficientNet (balanced)

Model-Specific Preprocessing:

• Each pre-trained model expects specific input format

• Always use model’s preprocessing function (e.g., mobilenet_v2.preprocess_input)

• MobileNetV2 scales to [-1, 1], ResNet subtracts mean, EfficientNet normalizes

Feature Extraction:

• Freeze pre-trained backbone (trainable=False)

• Train only new classification layers

• Fast training, provides good baseline performance

Hybrid Approach:

• Combine custom CNN branch with pre-trained features

• Learn task-specific patterns while leveraging ImageNet knowledge

• Useful when dataset has unique characteristics

Fine-tuning:

• Unfreeze backbone after initial training with frozen weights

• Continue training with very low learning rate (e.g., 1e-5 vs default 1e-3)

• Allows weights to adapt to specific task without forgetting ImageNet features

• Usually improves accuracy by 2-5% over frozen features

Best Practices:

• Always start with frozen backbone training

• Use correct preprocessing for chosen backbone

• Use very low learning rate for fine-tuning (100× smaller)

• Monitor validation metrics to avoid overfitting