TFhub、Keras预训练模型进行推理简单示例

最新推荐文章于 2025-05-15 00:24:48 发布

fwu11

最新推荐文章于 2025-05-15 00:24:48 发布

阅读量1.2k

点赞数 1

分类专栏： Tensorflow

本文链接：https://blog.youkuaiyun.com/weixin_42552135/article/details/110375235

版权

导入预训练模型进行推理。以mobilenetv2 Imagenet预训练模型为例：

import tensorflow as tf
import tensorflow_hub as hub
import numpy as np

# 加载图片
image = tf.image.decode_image(tf.io.read_file('/path/to/your/image'),channels=3)
image = tf.image.resize(image,(224,224))
image = tf.expand_dims(image, axis=0)

# Keras预测
i = tf.keras.layers.Input([224, 224, 3], dtype = tf.uint8)
x = tf.cast(i, tf.float32)
x = tf.keras.applications.mobilenet.preprocess_input(x)
core = tf.keras.applications.MobileNetV2()
x = core(x)
model = tf.keras.Model(inputs=[i], outputs=[x])
preds = model(image).numpy()
print(tf.keras.applications.mobilenet_v2.decode_predictions(preds, top=5))

# TFHub预测
path = "https://tfhub.dev/google/tf2-preview/mobilenet_v2/classification/4"
IMAGE_SHAPE = (224, 224)
tfhub_model = tf.keras.Sequential([
    hub.KerasLayer(path, input_shape=IMAGE_SHAPE+(3,), outpu