MLflow transformers

This is an example of how MLflow transformers can be supported in Cloudera AI.

This example shows how to implement a translation workflow using a translation model.

1. Save the following as a file, for example, named mlflowtest.py.

   #!pip3 install torch transformers torchvision tensorflow
       
       import mlflow
       from mlflow.models import infer_signature
       from mlflow.transformers import generate_signature_output
       from transformers import pipeline
       
       en_to_de = pipeline("translation_en_to_de")
       
       data = "MLflow is great!"
       output = generate_signature_output(en_to_de, data)
       #signature = infer_signature(data, output)
       
       with mlflow.start_run() as run:
       mlflow.transformers.log_model(
       transformers_model=en_to_de,
       artifact_path="english_to_german_translator",
       input_example=data,
       registered_model_name="entodetranslator",
       )
       
       
       model_uri = f"runs:/{run.info.run_id}/english_to_german_translator"
       loaded = mlflow.pyfunc.load_model(model_uri)
       
       print(loaded.predict(data))

2. In the AI Registry page, find the entodetranslator model. Deploy the model.
3. Make a request using the following payload:

   {
     "dataframe_split": {
       "columns": [
         "data"
       ],
       "data": [
         [
           "MLflow is great!"
         ]
       ]
     }
   }

4. In a session, run the mlflowtest.py file. It should print the following output.

   print(loaded.predict(data))
   
   ['MLflow ist großartig!']