After using mlflow to deploy the model into docker, input data shows error.

[lyw615]

System information

The model is deployed into docker which is from mlflow/examples/sklearn_elasticnet_wine/train.py. Following are details of docker.
docker run -p 5001:8080 "ni"
pip 19.0.3 from /miniconda/lib/python3.7/site-packages/pip (python 3.7)
Python 3.7.3
1.0.0
[2019-07-24 11:55:06 +0000] [44] [INFO] Starting gunicorn 19.9.0
[2019-07-24 11:55:06 +0000] [44] [INFO] Listening at: unix:/tmp/gunicorn.sock (44)
[2019-07-24 11:55:06 +0000] [44] [INFO] Using worker: gevent
[2019-07-24 11:55:06 +0000] [47] [INFO] Booting worker with pid: 47

127.0.0.1 - - [24/Jul/2019:12:12:51 +0000] "POST /invocations HTTP/1.1" 400 1174 "-" "curl/7.47.0"

Describe the problem

Inside docker enter data with the following command :
curl http://127.0.0.1:8080/invocations -H 'Content-Type: application/json' -d '{ "columns": ["a", "b", "c"], "data": [[1, 2, 3], [4, 5, 6]] }'
Error report:
{"error_code": "BAD_REQUEST", "message": "Encountered an unexpected error while evaluating the model. Verify that the serialized input Dataframe is compatible with the model for inference.", "stack_trace": "Traceback (most recent call last):\n File \"/miniconda/envs/custom_env/lib/python3.6/site-packages/mlflow/pyfunc/scoring_server/__init__.py\", line 196, in transformation\n raw_predictions = model.predict(data)\n File \"/miniconda/envs/custom_env/lib/python3.6/site-packages/sklearn/linear_model/base.py\", line 256, in predict\n return self._decision_function(X)\n File \"/miniconda/envs/custom_env/lib/python3.6/site-packages/sklearn/linear_model/coordinate_descent.py\", line 791, in _decision_function\n return super(ElasticNet, self)._decision_function(X)\n File \"/miniconda/envs/custom_env/lib/python3.6/site-packages/sklearn/linear_model/base.py\", line 241, in _decision_function\n dense_output=True) + self.intercept_\n File \"/miniconda/envs/custom_env/lib/python3.6/site-packages/sklearn/utils/extmath.py\", line 140, in safe_sparse_dot\n return np.dot(a, b)\nValueError: shapes (2,3) and (11,) not aligned: 3 (dim 1) != 11 (dim 0)\n"}

[lyw615]

I know that there may be a problem with the format of my data input, but I can't find other information.

[lyw615]

My input was error,sorry very much

[vijayakumarmariyappan]

@lyw615 Me too facing the same problem.
Your input is also fine right
curl http://127.0.0.1:8080/invocations -H 'Content-Type: application/json' -d '{ "columns": ["a", "b", "c"], "data": [[1, 2, 3], [4, 5, 6]] }'

How did you fix this? meant what changes you did for the above url?

THanks for helping :)

[Yeshvini]

Could you @lyw615 please tell , how you fixed this issue?

[vatsaRJOYO]

I got the same issue as well. To aid in type inferencing, i had to log model with model signature. Now, deploy the new model saved with model signature, It then works like a charm.
reference: https://www.mlflow.org/docs/latest/models.html#how-to-log-models-with-signatures