Description
Pandas UDF is the ideal connection between PySpark and DL model inference workload. However, user needs to load the model file first to make predictions. It is common to see models of size ~100MB or bigger. If the Pandas UDF execution is limited to each batch, user needs to repeatedly load the same model for every batch in the same python worker process, which is inefficient.
We can provide users the iterator of batches in pd.DataFrame and let user code handle it:
@pandas_udf(DoubleType(), PandasUDFType.SCALAR_ITER)
def predict(batch_iter):
model = ... # load model
for batch in batch_iter:
yield model.predict(batch)
The type of each batch is:
- a pd.Series if UDF is called with a single non-struct-type column
- a tuple of pd.Series if UDF is called with more than one Spark DF columns
- a pd.DataFrame if UDF is called with a single StructType column
Examples:
@pandas_udf(...) def evaluate(batch_iter): model = ... # load model for features, label in batch_iter: pred = model.predict(features) yield (pred - label).abs() df.select(evaluate(col("features"), col("label")).alias("err"))
@pandas_udf(...) def evaluate(pdf_iter): model = ... # load model for pdf in pdf_iter: pred = model.predict(pdf['x']) yield (pred - pdf['y']).abs() df.select(evaluate(struct(col("features"), col("label"))).alias("err"))
If the UDF doesn't return the same number of records for the entire partition, user should see an error. We don't restrict that every yield should match the input batch size.
Another benefit is with iterator interface and asyncio from Python, it is flexible for users to implement data pipelining.
cc: icexelloss bryanc holdenk hyukjin.kwon ueshin smilegator
Attachments
Issue Links
- blocks
-
SPARK-28056 Document SCALAR_ITER Pandas UDF
- Resolved
- is related to
-
SPARK-24579 SPIP: Standardize Optimized Data Exchange between Spark and DL/AI frameworks
- Open
- relates to
-
SPARK-26413 SPIP: RDD Arrow Support in Spark Core and PySpark
- Open
- links to