ValueError：训练模型时YOLO

问题描述

在Google Colab上训练模型时，我遇到以下错误，亚当依赖项/优化器可能存在问题。
ValueError：张量转换请求具有dtype float32的Tensor的dtype float32_ref：
有什么办法可以解决此问题？
训练脚本的代码：
import os
import sys
import argparse

def get_parent_dir(n=1):
    """ returns the n-th parent dicrectory of the current
    working directory """
    current_path = os.path.dirname(os.path.abspath(__file__))
    for k in range(n):
        current_path = os.path.dirname(current_path)
    return current_path

src_path = os.path.join(get_parent_dir(0),'src')
sys.path.append(src_path)

utils_path = os.path.join(get_parent_dir(1),'Utils')
sys.path.append(utils_path)

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 keras_yolo3.yolo3.model import preprocess_true_Boxes,yolo_body,tiny_yolo_body,yolo_loss
from keras_yolo3.yolo3.utils import get_random_data
from PIL import Image
from time import time
import pickle

from Train_Utils import get_classes,get_anchors,create_model,create_tiny_model,data_generator,data_generator_wrapper,ChangetoOtherMachine


keras_path = os.path.join(src_path,'keras_yolo3')
Data_Folder = os.path.join(get_parent_dir(1),'Data')
Image_Folder = os.path.join(Data_Folder,'Source_Images','Training_Images')
Vott_Folder = os.path.join(Image_Folder,'vott-csv-export')
YOLO_filename = os.path.join(Vott_Folder,'data_train.txt')

Model_Folder = os.path.join(Data_Folder,'Model_Weights')
YOLO_classname = os.path.join(Model_Folder,'data_classes.txt')

log_dir = Model_Folder
anchors_path = os.path.join(keras_path,'model_data','yolo_anchors.txt') 
weights_path = os.path.join(keras_path,'yolo.h5') 

FLAGS = None

if __name__ == '__main__':
    # Delete all default flags
    parser = argparse.ArgumentParser(argument_default=argparse.SUPPRESS)
    '''
    Command line options
    '''

    parser.add_argument(
        "--annotation_file",type=str,default=YOLO_filename,help = "Path to annotation file for Yolo. Default is "+ YOLO_filename
        )
    parser.add_argument(
        "--classes_file",default=YOLO_classname,help = "Path to YOLO classnames. Default is "+ YOLO_classname
        )

    parser.add_argument(
        "--log_dir",default=log_dir,help = "Folder to save training logs and trained weights to. Default is "+ log_dir 
        )

    parser.add_argument(
        "--anchors_path",default=anchors_path,help = "Path to YOLO anchors. Default is "+ anchors_path
        )

    parser.add_argument(
        "--weights_path",default=weights_path,help = "Path to pre-trained YOLO weights. Default is "+ weights_path
        )
    parser.add_argument(
        "--val_split",type=float,default=0.1,help = "Percentage of training set to be used for validation. Default is 10%."
        )
    parser.add_argument(
        "--is_tiny",default=False,action="store_true",help = "Use the tiny Yolo version for better performance and less accuracy. Default is False."
        )
    parser.add_argument(
        "--random_seed",default=None,help = "Random seed value to make script deterministic. Default is 'None',i.e. non-deterministic."
        )
    parser.add_argument(
        "--epochs",default=51,help = "Number of epochs for training last layers and number of epochs for fine-tuning layers. Default is 51."
        )

    
    FLAGS = parser.parse_args()

    np.random.seed(FLAGS.random_seed)

    log_dir = FLAGS.log_dir

    class_names = get_classes(FLAGS.classes_file)
    num_classes = len(class_names)
    anchors = get_anchors(FLAGS.anchors_path)
    weights_path = FLAGS.weights_path

    input_shape = (416,416) # multiple of 32,height,width
    epoch1,epoch2 = FLAGS.epochs,FLAGS.epochs

    is_tiny_version = (len(anchors)==6) # default setting
    if FLAGS.is_tiny:
        model = create_tiny_model(input_shape,anchors,num_classes,freeze_body=2,weights_path = weights_path)
    else:
        model = create_model(input_shape,weights_path = weights_path) # make sure you kNow what you freeze

    log_dir_time = os.path.join(log_dir,'{}'.format(int(time())))
    logging = TensorBoard(log_dir=log_dir_time)
    checkpoint = ModelCheckpoint(os.path.join(log_dir,'checkpoint.h5'),monitor='val_loss',save_weights_only=True,save_best_only=True,period=5)
    reduce_lr = ReduceLROnPlateau(monitor='val_loss',factor=0.1,patience=3,verbose=1)
    early_stopping = EarlyStopping(monitor='val_loss',min_delta=0,patience=10,verbose=1)

    val_split = FLAGS.val_split
    with open(FLAGS.annotation_file) as f:
        lines = f.readlines()

    # This step makes sure that the path names correspond to the local machine 
    # This is important if annotation and training are done on different machines (e.g. training on AWS)
    lines  = ChangetoOtherMachine(lines,remote_machine = '')
    np.random.shuffle(lines)
    num_val = int(len(lines)*val_split)
    num_train = len(lines) - num_val

    # Train with frozen layers first,to get a stable loss. 
    # Adjust num epochs to your dataset. This step is enough to obtain a decent model.
    if True:
        model.compile(optimizer=Adam(lr=1e-3),loss={
            # use custom yolo_loss Lambda layer.
            'yolo_loss': lambda y_true,y_pred: y_pred})

        batch_size = 32
        print('Train on {} samples,val on {} samples,with batch size {}.'.format(num_train,num_val,batch_size))
        history = model.fit_generator(data_generator_wrapper(lines[:num_train],batch_size,input_shape,num_classes),steps_per_epoch=max(1,num_train//batch_size),validation_data=data_generator_wrapper(lines[num_train:],validation_steps=max(1,num_val//batch_size),epochs=epoch1,initial_epoch=0,callbacks=[logging,checkpoint])
        model.save_weights(os.path.join(log_dir,'trained_weights_stage_1.h5'))

        step1_train_loss = history.history['loss']
        
        file = open(os.path.join(log_dir_time,'step1_loss.npy'),"w")
        with open(os.path.join(log_dir_time,'w') as f:
            for item in step1_train_loss:
                f.write("%s\n" % item) 
        file.close()
        
        step1_val_loss = np.array(history.history['val_loss'])
        
        file = open(os.path.join(log_dir_time,'step1_val_loss.npy'),'w') as f:
            for item in step1_val_loss:
                f.write("%s\n" % item) 
        file.close()
        
    # Unfreeze and continue training,to fine-tune.
    # Train longer if the result is unsatisfactory.
    if True:
        for i in range(len(model.layers)):
            model.layers[i].trainable = True
        model.compile(optimizer=Adam(lr=1e-4),loss={'yolo_loss': lambda y_true,y_pred: y_pred}) # recompile to apply the change
        print('Unfreeze all layers.')

        batch_size = 4 # note that more GPU memory is required after unfreezing the body
        print('Train on {} samples,batch_size))
        history=model.fit_generator(data_generator_wrapper(lines[:num_train],epochs=epoch1+epoch2,initial_epoch=epoch1,checkpoint,reduce_lr,early_stopping])
        model.save_weights(os.path.join(log_dir,'trained_weights_final.h5'))
        step2_train_loss = history.history['loss']
        
        file = open(os.path.join(log_dir_time,'step2_loss.npy'),'w') as f:
            for item in step2_train_loss:
                f.write("%s\n" % item) 
        file.close()
        
        step2_val_loss = np.array(history.history['val_loss'])
        
        file = open(os.path.join(log_dir_time,'step2_val_loss.npy'),'w') as f:
            for item in step2_val_loss:
                f.write("%s\n" % item) 
        file.close()
解决方法

暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！
如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。
小编邮箱:dio#foxmail.com (将#修改为@）
computer-vision deep-learning keras tensorflow training-data
ValueError：训练模型时YOLO_v3

问题描述

解决方法
