版本信息:
PaddlePaddle1.7.2
PaddleHub1.6.0
Python3.7.4

环境:
Windows使用AIStudio

在PaddlePaddle1.7.2上进行finetune训练后, 预测结果里只出现自定义lablelist里的lable，resnet50_vd_animals模型网络结构被改变无法正常预测出原模型的lablelist, 请问这种情况应该怎么解决？

相关代码如下:

#加载自定义数据集
#!cp /home/aistudio/data/data103940/test.zip ./
#!unzip -o test.zip

import argparse
import os
import ast

import paddle.fluid as fluid
import paddlehub as hub
import numpy as np

from paddlehub.dataset.base_cv_dataset import BaseCVDataset

class DemoDataset(BaseCVDataset):
def __init__(self):
# 数据集存放位置
self.dataset_dir = "test"
super(DemoDataset, self).__init__(
base_path=self.dataset_dir,
train_list_file="train_list.txt",
validate_list_file="validate_list.txt",
test_list_file="test_list.txt",
#predict_file="predict_list.txt",
label_list_file="label_list.txt",
# label_list=["数据集所有类别"]
)
dataset = DemoDataset()
print("加载自定义数据集完成!")

module = hub.Module(name="resnet50_vd_animals")
print("加载预训练模型完成!")

#生成Reader
# Use ImageClassificationReader to read dataset
data_reader = hub.reader.ImageClassificationReader(
image_width=module.get_expected_image_width(),
image_height=module.get_expected_image_height(),
images_mean=module.get_pretrained_images_mean(),
images_std=module.get_pretrained_images_std(),
dataset=dataset)
print("生成Reader完成!")

#配置策略

# Setup runing config for PaddleHub Finetune API
config = hub.RunConfig(
use_cuda=True,
num_epoch=100,
checkpoint_dir="./cv_finetune_turtorial_demo",
batch_size=32,
eval_interval=50,
strategy=hub.finetune.strategy.DefaultFinetuneStrategy())
print("配置策略完成!")

#迁移组网
# Define a reading comprehension finetune task by PaddleHub's API
input_dict, output_dict, program = module.context(trainable=True) #获取预训练module的上下文信息包括输入、输出变量以及paddle program

feature_map = output_dict["feature_map"] #从预训练模型的输出变量中找到特征图，提取feature_map

feed_list = [input_dict["image"].name] #获取预训练模型中输入变量的名字列表，需要把该名字列表feed进入组建的task中

task = hub.ImageClassifierTask(
data_reader=data_reader,
feed_list=feed_list,
feature=feature_map,
num_classes=dataset.num_labels,
config=config)
print("迁移组网完成! 即将开始一个"+str(dataset.num_labels)+"分类任务。")

#迁移任务执行
task.finetune_and_eval()
print("finetuning 完成!")

import paddlehub as hub
import numpy as np

data = ["0_1152.jpg"]
label_map = dataset.label_dict()
index = 0
# get classification result
run_states = task.predict(data=data)
results = [run_state.run_results for run_state in run_states]

for batch_result in results:
# get predict index
batch_result = np.argmax(batch_result, axis=2)[0]
for result in batch_result:
index += 1
result = label_map[result]
print("input %i is %s, and the predict result is %s" %
(index, data[index - 1], result))