Cannot load the model of the PneumoniaDetection tuto

[stevo32800]

I am trying to follow the PneumonialDetection tuto but i encountered an issue. I have a problem with the input/output i don't know how to solve this problem.

0 [main] INFO net.PneumoniaDetection  - Input image not specified, using image:
	https://djl-ai.s3.amazonaws.com/resources/images/chest_xray.jpg
1323 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - Loading model with Criteria:
	Application: UNDEFINED
	Input: interface ai.djl.modality.cv.Image
	Output: class ai.djl.modality.Classifications

1347 [main] DEBUG ai.djl.engine.Engine  - No engine found from EngineProvider
1475 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - Checking ModelLoader: ai.djl.zoo:mlp CV.IMAGE_CLASSIFICATION [
	ai.djl.zoo/mlp/0.0.3/mlp {"dataset":"mnist"}
]
1482 [main] DEBUG ai.djl.repository.MRL  - Preparing artifact: zoo, ai.djl.zoo/mlp/0.0.3/mlp {"dataset":"mnist"}
1485 [main] DEBUG ai.djl.repository.AbstractRepository  - Files have been downloaded already: C:\Users\a871169\.djl.ai\cache\repo\model\cv\image_classification\ai\djl\zoo\mlp\mnist\0.0.3
1488 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - No supported engine available for model zoo: ai.djl.zoo for ModelLoader: ai.djl.zoo:mlp
1493 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - Checking ModelLoader: ai.djl.zoo:resnet CV.IMAGE_CLASSIFICATION [
	ai.djl.zoo/resnet/0.0.2/resnetv1 {"layers":"50","flavor":"v1","dataset":"cifar10"}
]
1494 [main] DEBUG ai.djl.repository.MRL  - Preparing artifact: zoo, ai.djl.zoo/resnet/0.0.2/resnetv1 {"layers":"50","flavor":"v1","dataset":"cifar10"}
1496 [main] DEBUG ai.djl.repository.AbstractRepository  - Files have been downloaded already: C:\Users\a871169\.djl.ai\cache\repo\model\cv\image_classification\ai\djl\zoo\resnet\50\v1\cifar10\0.0.2
1498 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - No supported engine available for model zoo: ai.djl.zoo for ModelLoader: ai.djl.zoo:resnet
1503 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - Checking ModelLoader: ai.djl.zoo:ssd CV.OBJECT_DETECTION [
	ai.djl.zoo/ssd/0.0.2/ssd {"flavor":"tiny","dataset":"pikachu"}
]
1503 [main] DEBUG ai.djl.repository.MRL  - Preparing artifact: zoo, ai.djl.zoo/ssd/0.0.2/ssd {"flavor":"tiny","dataset":"pikachu"}
1504 [main] DEBUG ai.djl.repository.AbstractRepository  - Files have been downloaded already: C:\Users\a871169\.djl.ai\cache\repo\model\cv\object_detection\ai\djl\zoo\ssd\tiny\pikachu\0.0.2
1505 [main] DEBUG ai.djl.repository.zoo.ModelZoo  - No supported engine available for model zoo: ai.djl.zoo for ModelLoader: ai.djl.zoo:ssd
Exception in thread "main" ai.djl.repository.zoo.ModelNotFoundException: No matching model with specified Input/Output type found.
	at ai.djl.repository.zoo.Criteria.loadModel(Criteria.java:180)
	at net.PneumoniaDetection.main(PneumoniaDetection.java:67)
Caused by: ai.djl.repository.zoo.ModelNotFoundException: No supported engine available for model zoo: ai.djl.zoo
	at ai.djl.repository.zoo.BaseModelLoader.loadModel(BaseModelLoader.java:135)
	at ai.djl.repository.zoo.Criteria.loadModel(Criteria.java:168)
	... 1 more

Process finished with exit code 1

public class PneumoniaDetection {

    private static final Logger logger = LoggerFactory.getLogger(PneumoniaDetection.class);

    private static final List<String> CLASSES = Arrays.asList("Normal", "Pneumonia");

    public static void main(String[] args) throws IOException, TranslateException, ModelException {
        String imagePath;
        BasicConfigurator.configure();
        if (args.length == 0) {
            imagePath = "https://djl-ai.s3.amazonaws.com/resources/images/chest_xray.jpg";
            logger.info("Input image not specified, using image:\n\t{}", imagePath);
        } else {
            imagePath = args[0];
        }
        Image image;
        if (imagePath.startsWith("http")) {
            image = ImageFactory.getInstance().fromUrl(imagePath);
        } else {
            image = ImageFactory.getInstance().fromFile(Paths.get(imagePath));
        }

        Translator<Image, Classifications> translator =
                ImageClassificationTranslator.builder()
                        .addTransform(a -> NDImageUtils.resize(a, 224).div(255.0f))
                        .optSynset(CLASSES)
                        .build();
        Criteria<Image, Classifications> criteria =
                Criteria.builder()
                        .setTypes(Image.class, Classifications.class)
                        .optModelName("saved_model")
                        .optTranslator(translator)
                        .build();

        try (ZooModel<Image, Classifications> model = criteria.loadModel();
             Predictor<Image, Classifications> predictor = model.newPredictor()) {
            Classifications result = predictor.predict(image);
            logger.info("Diagnose: {}", result);
        }


    }
}

[frankfliu]

You need download pre-trained model manually: https://github.com/deepjavalibrary/djl-demo/tree/master/pneumonia-detection#prepare-trained-model

[stevo32800]

Thanks for your answer, it's actually what i did. But in which directory should i put the model ?

[frankfliu]

The model should be downloaded in djl-demo/pneumonia-detection/models folder.

I created a PR to download the model automatically if you didn't do it manually, please get latest code: deepjavalibrary/djl-demo#292

[stevo32800]

hello, I did what you said, i put the model in the models folder in the root of the pneumonia-detection example and i still have the same error. I don't undestand what can be wrong... the is a method to give the path of the model ?

[frankfliu]

@stevo32800
Can you try to run command line in console: ./gradlew?

Your output looks suspicious, are you running IDE?

1. there are numbers in each line, not sure where its comes from
2. By default in our gradle settings the debug log is not enabled, and your out has debug output
3. We didn't include model-zoo package in gradle dependencies, why it is included?
4. Did you include tensorflow dependency in your project?