BentoML

```f=open('extra.pkl', 'rb')
model = pickle.load(f)
f.close()
# saved_model = bentoml.sklearn.save_model("extra", model)
# print(f"Model saved: {saved_model}")
extra= bentoml.sklearn.get('extra:latest').to_runner()```

here is my code to load model from pickle.

```    result_extra = extra.predict.run([output])[0]```

after loadmodel. here is my code to predict

yes. I do exactly the same as randomforest, it's ok

Are you calling the endpoint from an HTTP client?

yes. i got a bug in the predict step. The randomforest model is ok, so I wonder if bentoml can load the extratree model or not?

Hm, have you passed that runner into the service?

yes. I got bug in predict step. "    result_extra = extra.predict.run([output])[0]"

Right; can you send me the line where you define your service?

In the line where you do `bentoml.Service` you need to pass `extra` along with `[rf,xgb,extra]`

:+1: no problem, that error should definitely be made more clear