-
After creating a PySpark model and serializing it to a bundle, I try to read in the MLeap transformer and make a prediction but the prediction is wrong.
Upon investigation, I've found that the inpu…
-
The OneVsRest Classifier Spark model (https://spark.apache.org/docs/2.4.0/ml-classification-regression.html#one-vs-rest-classifier-aka-one-vs-all) is not supported by Splice as it cannot be serialized…
-
Hi, I've already asked the [question on StackOverflow](https://stackoverflow.com/questions/54126259/is-mleap-actually-a-serialization-format), but let me duplicate it here.
I began to work with [ML…
-
import org.apache.spark.ml.feature.OneHotEncoderEstimator
depedency : "ml.combust.mleap" %% "mleap-spark" % "0.15.0"
@hollinwilkins tried your suggestion from similar issue, still getting the sa…
-
Hi,
Props to the creators and maintainers of the mleap library. I have used it for run time and it's serving speed is blazing fast. I ran in to one error while implementing mleap library for our us…
-
I have written a basic script with pyspark for writing a serialized model to HDFS. Here's the code:
```import mleap.pyspark
from mleap.pyspark.spark_support import SimpleSparkSerializer
from pyspar…
-
I am training a CountVectorizerModel in Spark, where it takes an array of strings as a input feature value. I then serialized the model using MLeap, and deserialize it in a pure Java environment and u…
-
I don't see that this is being serialized, and an exception occurs on mleap side when unseen data happens in leapframe.
mtsol updated
4 years ago
-
it seems that this bug not fixed for 0.11 version with this case: #489
-
### System information
- **Have I written custom code (as opposed to using a stock example script provided in MLflow)**: No
- **OS Platform and Distribution (e.g., Linux Ubuntu 16.04)**: Mac OSX
- …