keras-yolo3项目之训练文件train.py注释

最新推荐文章于 2024-05-29 18:06:44 发布

great-wind

最新推荐文章于 2024-05-29 18:06:44 发布

阅读量212

点赞数

分类专栏：深度学习

本文链接：https://blog.csdn.net/csdn1e/article/details/114985446

版权

本文详细注释了keras-yolo3项目的训练文件train.py，通过阅读可以深入理解模型的训练流程和逻辑。通过训练代码，你可以掌握YOLOv3在Keras中的实现细节。

摘要由CSDN通过智能技术生成

要想知道一个模型的流程，个人认为，看训练代码是能最快了解模型前后逻辑关系的方法之一，keras-yolo3项目训练源码如下：

"""
Retrain the YOLO model for your own dataset.
使用自己的数据训练YOLO模型
"""

import numpy as np
import keras.backend as K
from keras.layers import Input, Lambda
from keras.models import Model
from keras.optimizers import Adam
from keras.callbacks import TensorBoard, ModelCheckpoint, ReduceLROnPlateau, EarlyStopping

from yolo3.model import preprocess_true_boxes, yolo_body, tiny_yolo_body, yolo_loss
from yolo3.utils import get_random_data

import tensorflow as tf


def _main():
    annotation_path = '2007_train.txt'
    log_dir = 'logs/000/'
    classes_path = 'model_data/smoking_classes.txt'
    anchors_path = 'model_data/smoking_anchors.txt'
    class_names = get_classes(classes_path)
    num_classes = len(class_names)
    anchors = get_anchors(anchors_path)

    input_shape = (416,416) # multiple of 32, hw

    is_tiny_version = len(anchors)==6 # default setting
    if is_tiny_version:
        model = create_tiny_model(input_shape, anchors, num_classes,
            freeze_body=2, weights_path='model_data/tiny_yolo_weights.h5')    #构建初始模型,即模型初始化
    else:
        # 创建初始模型，并冻结两层网络层
        model = create_model(input_shape, anchors, num_classes,
            freeze_body=2, weights_path='model_data/yolo_weights.h5') # make sure you know what you freeze

    logging = TensorBoard(log_dir=log_dir)
    checkpoint = ModelCheckpoint(log_dir + 'ep{epoch:03d}-loss{loss:.3f}-val_loss{val_loss:.3f}.h5',
        monitor='val_loss', save_weights_only=True, save_best_only=True, period=3)    #保存权重
    reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=3, verbose=1)    #学习率
    early_stopping = EarlyStopping(monitor='val_loss', min_delta=0, patience=100, verbose=1)   #早停设置

    val_split = 0.1    #验证集比重
    # 读取标注数据
    with open(annotation_path) as f:
        lines = f.readlines()
    np.random.seed(10101)
    np.random.shuffle(lines)
    np.random.seed(None)
    num_val = int(len(lines)*val_split)    # 验证集数量
    num_train = len(lines) - num_val    #训练集数量

    # Train with frozen layers first, to get a stable loss.
    # Adjust num epochs to your dataset. This step is enough to obtain a not bad model.
    if True