This message was deleted.

Hey! Welcome to BentoML, how can we help

On the page that I linked, I am specifically following the instructions under heading: Pre-built Model Runners. When I run this code I get an error that states: AttributeError: 'Model' object has no attribute 'predict'

can you share the code that you are running?

you can simply run

runner.predict.run(..)

. this is analogous to doing

model(..)

runner = bentoml.pytorch.get("pytorch_model:xzrlgzsovglrsycf").to_runner() runner.init_local() print(runner.predict.run([[133.171875, 59.716081, 0.043133, -0.703383, 54.917224, 70.084438, 0.749798, -0.649512]]))

SORRY

for some reason that is still not working. do you think it has to do with how I saved my model? should I save it like:

bentoml.pytorch.save(
    model,
    "my_torch_model",
    signatures={"__call__": {"batchable": True, "batch_dim": 0}},
)

instead of like:

bentoml.pytorch.save_model(
    "demo_mnist",  # model name in the local model store
    trained_model,  # model instance being saved
    signatures={   # model signatures for runner inference
        "predict": {
            "batchable": True,
            "batch_dim": 0,
        }
    }
)

you are right,

signatures={   # model signatures for runner inference
        "predict": {
            "batchable": True,
            "batch_dim": 0,
        }
    }

this expects a

predict

function in your model. for pytorch model you could see this page https://docs.bentoml.org/en/latest/frameworks/pytorch.html

I appreciate your help!!!!

you're welcome! this also points out that we probably need to more clear in our documentation!