BentoML

```pre_model = bentoml.mlflow.import_model(
    model_name,
    model_uri,
    signatures={'predict': {'batchable': True}}   &lt;-----
)```

```pre_model = bentoml.mlflow.import_model(
    model_name,
    model_uri,
    signatures={'nhwc': {'batchable': False}}
)```


Thanks Jiang but I got...
`MLflow pyfunc model support only the predict method, signatures={'nhwc': {'batchable': False}} is not supported`

I think torch version of the model and weights is the problem.
• So I model training and update weight with `torch 1.13.0+cu117`
• retrieve model with `torch 1.13`.
• Done well!