标题翻译
Integrate Google Teachable Machine Model to Java application
问题
我在 https://teachablemachine.withgoogle.com 上创建了一个图像模型，用于将简单的标准化图标分类，因为我在我的应用程序中需要这个功能，但我未能将其导入到我的 Java 应用程序中。我尝试了从 .h5 导入，通过 Deeplearning4j（不受支持）到与 Tensorflow 和 savedModel 的奋斗。其他导出格式包括 Tensorflow.js 和 Tensorflow Lite。
我尝试在 Python 中手动训练模型，但由于我目前的深度学习技能有限，这个过程过于复杂，而且效果也不如 Teachable Machine 好。
我的程序必须从图像中对已知的GHS图标进行分类（https://en.wikipedia.org/wiki/Globally_Harmonized_System_of_Classification_and_Labelling_of_Chemicals）。
我该怎么办？是否有另一种方法可以集成模型，或者在这种简单情况下是否有比深度学习更简单的方法？
英文翻译
I made an image model with https://teachablemachine.withgoogle.com, which classifies simple
standardized pictograms as I need this in my application but I didn't manage to import it to my java application. I tried everything from .h5 import via Deeplearning4j (unsupported) to fighting around with Tensorflow and savedModel. Other export-formats are Tensorflow.js and Tensorflow Lite.
My attempt to train a model manually in python was too complex for my current deeplearning skills and never worked as good as the teachable machine.
My program has to classify the known GHS-pictograms from images (https://en.wikipedia.org/wiki/Globally_Harmonized_System_of_Classification_and_Labelling_of_Chemicals)
What should I do? Is there another way to integrate the model or maybe is there a simpler way than deeplearning for this simple thing?
答案1
得分: 0
首先，将您的模型导出为SavedModel格式。
然后您可以使用机器学习库来读取它。
例如：
示例代码（取自此处）：
Criteria<Image, Classifications> criteria =
Criteria.builder()
.setTypes(Image.class, Classifications.class)
.optModelUrls("https://example.com/squeezenet.zip")
.optTranslator(ImageClassificationTranslator
.builder().addTransform(new ToTensor()).build())
.build();
try (ZooModel<Image, Classification> model = ModelZoo.load(criteria);
Predictor<Image, Classification> predictor = model.newPredictor()) {
Image image = ImageFactory.getInstance().fromUrl("https://myimage.jpg");
Classification result = predictor.predict(image);
}
英文翻译
First Export your model as SavedModel.
Then you can use ML Libraries to read it.
For Example:-
Example code (Taken from here):-
Criteria<Image, Classifications> criteria =
Criteria.builder()
.setTypes(Image.class, Classifications.class)
.optModelUrls("https://example.com/squeezenet.zip")
.optTranslator(ImageClassificationTranslator
.builder().addTransform(new ToTensor()).build())
.build();
try (ZooModel<Image, Classification> model = ModelZoo.load(criteria);
Predictor<Image, Classification> predictor = model.newPredictor()) {
Image image = ImageFactory.getInstance().fromUrl("https://myimage.jpg");
Classification result = predictor.predict(image);
}
专注分享java语言的经验与见解，让所有开发者获益！
评论