topic Re: Performance issue while calling mlflow endpoint in Get Started Discussions

Performance issue while calling mlflow endpoint

sanjay — Sun, 28 Jan 2024 19:24:38 GMT

Hi,

I have pyspark dataframe and pyspark udf which calls mlflow model for each row but its performance is too slow.

Here is sample code

def myfunc(input_text):
restult = mlflowmodel.predict(input_text)
return result

myfuncUDF = udf(myfunc,StringType())

df = spark.sql("select * from test")
df=df.withColumn("test_result",myfuncUDF("input_text"))

Please suggest how to improve the performance.

Regards,
Sanjay

Re: Performance issue while calling mlflow endpoint

sanjay — Wed, 31 Jan 2024 12:28:29 GMT

Thank you Kaniz for the suggestions. This is really helpful. I even tried using applyInPandas. Not sure if this is better than spark UDF. If not can you help me in converting this function to pandas udf or any other optimized function.

def myfunc(input_text):

restult = mlflowmodel.predict(input_text)

return result

def myfuncUDF(pdf):

pdf['test_result']=pdf["input_text"].apply(myfunc)

return pdf

df = spark.sql("select * from test")

df = df.groupBy("id").applyInPandas(myfuncUDF)

Regards,

Sanjay

Re: Performance issue while calling mlflow endpoint

sanjay — Wed, 07 Feb 2024 07:36:39 GMT

Thank you @Retired_mod, its really helpful and did worked. Another quick question, I have to pass 2 parameters as input to myfunc. Please help how to pass multiple parameters.

def myfunc(input_text, param2):
# Assuming mlflowmodel is defined elsewhere
result = mlflowmodel.predict(input_text, param2)
return result

# Create a Pandas UDF
@pandas_udf(StringType())
def myfunc_udf(input_text_series: pd.Series, param2_series: pd.Series) -> pd.Series:
return input_text_series.apply(myfunc) ??

Re: Performance issue while calling mlflow endpoint

Isabeente — Wed, 07 Feb 2024 09:40:08 GMT

I need to send two arguments to myfunc, thus I have another brief question. I need some guidance on how to pass in many parameters.

Re: Performance issue while calling mlflow endpoint

sanjay — Wed, 07 Feb 2024 13:05:39 GMT

Hi Kaniz,

I started getting following error after using myfunc_udf with 2 parameters.

pythonException: 'ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all()

Regards,

Sanjay

Re: Performance issue while calling mlflow endpoint

BR_DatabricksAI — Wed, 07 Feb 2024 14:12:19 GMT

Hello Sanjay,

Could you please share your code snippet as per latest changes?

Re: Performance issue while calling mlflow endpoint

sanjay — Wed, 07 Feb 2024 15:39:45 GMT

Hi,

Here is code snippet as per latest changes.

def myfunc(t1,t2)

return 'test'

@pandas_udf(psf.StringType())

def myfunc_udf(input_text_series: pd.Series, param2_series: pd.Series) -> pd.Series:

def apply_myfunc(input_text, param2)

return myfunc(input_text, param2)

return input_text_series.apply(apply_myfunc, param2_series)

df.withColumn("result", myfunc_udf("input_text1", "input_text2"))

But I am getting error while running this

pythonException: 'ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all()

Re: Performance issue while calling mlflow endpoint

BR_DatabricksAI — Thu, 08 Feb 2024 05:25:35 GMT

Hello Sanjay,

The above code don't have the df defined. Can you share your df.show() output.

Re: Performance issue while calling mlflow endpoint

sanjay — Wed, 14 Feb 2024 14:42:28 GMT

Hi @Retired_mod Appreciate if you can help in resolving this issue.

Regards,

Sanjay

Re: Performance issue while calling mlflow endpoint

Isabeente — Tue, 19 Mar 2024 02:19:56 GMT

So good