Question

我正在尝试部署一个自定义的预测例程，该例程对文本进行分类。我训练了一个模型，并在云存储模型目录中提供了.pkl。模型版本创建失败，原因：

ERROR: (gcloud.beta.ai-platform.versions.create) Bad model detected with error:  "Failed to load model: User-provided package text_categorizer-0.2.tar.gz failed to install: Command '
['python-default', '-m', 'pip', 'install', '--target=/tmp/custom_lib', '--no-cache-dir', '-b', '/tmp/pip_builds', '/tmp/custom_code/text_categorizer-0.2.tar.gz']' returned non-zero e
xit status 1 (Error code: 0)"

还有什么日志可以检查，以了解更多关于为什么无法加载模型的信息？ Predictor类在上传到云存储中的.tar.gz文件中均可用。确保存储桶与创建模型的区域相同。模型文件位于名为model的目录中，而tar.gz文件位于名为textcateg的主目录下。 tar.gz文件包含预测变量python类。

创建版本的命令：

gcloud beta ai-platform versions create ver3 --model text_categorizer --runtime-version 1.13 --python-version 3.5 --origin gs://mdmceml-bucket/textcateg/model/ --package-uris gs://mdmceml-bucket/textcateg/text_categorizer-0.2.tar.gz --prediction-class predictor.GCPPredictor

GCPPredictor的相关代码段：：

    def __init__(self, le, clf):
        """Stores artifacts for prediction. Only initialized via `from_path`.
        """
        self._le = le
        self._clf = clf

    @classmethod
    def from_path(cls, model_dir):
        """Creates an instance of Predictor using the given path.

        Loading of the predictor should be done in this method.

        Args:
            model_dir: The local directory that contains the exported model
                file along with any additional files uploaded when creating the
                version resource.

        Returns:
            An instance implementing this Predictor class.
        """

        model_path = os.path.join(model_dir, 'category_classifier.pkl')
        with open(model_path, 'rb') as f:
            model = pickle.load(f)

        le_path = os.path.join(model_dir, 'labelencoder.pkl')
        with open(le_path, 'rb') as f1:
            le = pickle.load(f1)

        return cls(le, model)

创建模型版本失败，并显示以下信息：

ERROR: (gcloud.beta.ai-platform.versions.create) Bad model detected with error:  "Failed to load model: User-provided package text_categorizer-0.2.tar.gz failed to install: Command '
['python-default', '-m', 'pip', 'install', '--target=/tmp/custom_lib', '--no-cache-dir', '-b', '/tmp/pip_builds', '/tmp/custom_code/text_categorizer-0.2.tar.gz']' returned non-zero e
xit status 1 (Error code: 0)"

无法通过自定义预测例程部署经过训练的模型。因模型加载错误而失败

0 个答案: