使用模板在 Sagemaker 中部署 huggingface 零样本分类 returns 错误，缺少位置参数 'candidate_labels'

Question

我正在使用从 huggingface 生成的代码，任务：Zero-Shot Classification，配置：AWS 和运行在 Sagemaker 的 jupyterlab 中使用它

from sagemaker.huggingface import HuggingFaceModel
import sagemaker

role = sagemaker.get_execution_role()
# Hub Model configuration. https://huggingface.co/models
hub = {
    'HF_MODEL_ID':'facebook/bart-large-mnli',
    'HF_TASK':'zero-shot-classification'
}

# create Hugging Face Model Class
huggingface_model = HuggingFaceModel(
    transformers_version='4.6.1',
    pytorch_version='1.7.1',
    py_version='py36',
    env=hub,
    role=role, 
)

# deploy model to SageMaker Inference
predictor = huggingface_model.deploy(
    initial_instance_count=1, # number of instances
    instance_type='ml.m5.xlarge' # ec2 instance type
)

predictor.predict({
    'inputs': "Hi, I recently bought a device from your company but it is not working as advertised and I would like to get reimbursed!"
})

返回以下错误：

ModelError: An error occurred (ModelError) when calling the InvokeEndpoint operation: Received client error (400) from primary with message "{ "code": 400, "type": "InternalServerException",
"message": "call() missing 1 required positional argument: \u0027candidate_labels\u0027" } ". See ... in account **** for more information.

我试过运行以不同的方式设置它们，例如这样，

predictor.predict({
    'inputs': "Hi, I recently bought a device from your company but it is not working as advertised and I would like to get reimbursed!",
    'candidate_labels': ['science', 'life']
})

但还是不行。我应该怎么运行呢？

Answer 1

零样本分类模型的请求正文架构在此 link 中定义。

{
    "inputs": "Hi, I recently bought a device from your company but it is not working as advertised and I would like to get reimbursed!",
    "parameters": {
        "candidate_labels": [
            "refund",
            "legal",
            "faq"
        ]
    }
}

使用模板在 Sagemaker 中部署 huggingface 零样本分类 returns 错误，缺少位置参数 'candidate_labels'

Deploying huggingface zero-shot classification in Sagemaker using template returns error, missing positional argument 'candidate_labels'

python

model

amazon-sagemaker

huggingface-transformers