Skip to content

Commit

Permalink
Updated documentation
Browse files Browse the repository at this point in the history
  • Loading branch information
vruusmann committed Oct 31, 2019
1 parent 3f0ddee commit ab3805e
Showing 1 changed file with 33 additions and 33 deletions.
66 changes: 33 additions & 33 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -113,36 +113,36 @@ Evaluator evaluator = new ModelEvaluatorBuilder(javaPMML)

The effect of transpilation on memory consumption and execution speed can be estimated using the `org.jpmml.evaluator.EvaluationExample` command-line application (download the latest executable uber-JAR file from [JPMML-Evaluator releases](https://github.com/jpmml/jpmml-evaluator/releases) page).

For example, evaluating the `/src/test/resources/pmml/XGBoostAudit.pmml` model with the `/src/test/resources/csv/Audit.csv` input dataset.
For example, evaluating the `/src/test/resources/pmml/LightGBMAudit.pmml` model with the `/src/test/resources/csv/Audit.csv` input dataset.

Evaluating the model in interpreted mode:

```
$ java -jar pmml-evaluator-example-executable-${version}.jar --model XGBoostAudit.pmml --input Audit.csv --output /dev/null --optimize --intern --loop 100
$ java -jar pmml-evaluator-example-executable-${version}.jar --model LightGBMAudit.pmml --input Audit.csv --output /dev/null --optimize --intern --loop 100
-- Timers ----------------------------------------------------------------------
main
count = 1000
mean rate = 9,77 calls/second
1-minute rate = 9,36 calls/second
5-minute rate = 8,08 calls/second
15-minute rate = 7,65 calls/second
min = 94,77 milliseconds
max = 943,82 milliseconds
mean = 102,23 milliseconds
stddev = 30,24 milliseconds
median = 96,52 milliseconds
75% <= 99,77 milliseconds
95% <= 124,29 milliseconds
98% <= 153,55 milliseconds
99% <= 181,89 milliseconds
99.9% <= 943,14 milliseconds
mean rate = 3,52 calls/second
1-minute rate = 3,52 calls/second
5-minute rate = 3,16 calls/second
15-minute rate = 2,85 calls/second
min = 262,09 milliseconds
max = 1042,44 milliseconds
mean = 283,81 milliseconds
stddev = 39,11 milliseconds
median = 269,18 milliseconds
75% <= 280,05 milliseconds
95% <= 366,40 milliseconds
98% <= 373,75 milliseconds
99% <= 377,32 milliseconds
99.9% <= 1041,82 milliseconds
```

Transpiling the model:

```
$ java -jar target/jpmml-transpiler-executable-1.0-SNAPSHOT.jar --xml-input XGBoostAudit.pmml --jar-output XGBoostAudit.jar
$ java -jar target/jpmml-transpiler-executable-1.0-SNAPSHOT.jar --xml-input LightGBMAudit.pmml --jar-output LightGBMAudit.jar
```

Getting help:
Expand All @@ -154,28 +154,28 @@ $ java -jar target/jpmml-transpiler-executable-1.0-SNAPSHOT.jar --help
Evaluating the model in transpiled mode:

```
$ java -jar pmml-evaluator-example-executable-${version}.jar --model XGBoostAudit.jar --input Audit.csv --output /dev/null --loop 1000
$ java -jar pmml-evaluator-example-executable-${version}.jar --model LightGBMAudit.jar --input Audit.csv --output /dev/null --loop 1000
-- Timers ----------------------------------------------------------------------
main
count = 1000
mean rate = 77,61 calls/second
1-minute rate = 67,21 calls/second
5-minute rate = 66,09 calls/second
15-minute rate = 65,90 calls/second
min = 10,93 milliseconds
max = 239,73 milliseconds
mean = 12,82 milliseconds
stddev = 9,30 milliseconds
median = 11,05 milliseconds
75% <= 11,71 milliseconds
95% <= 16,54 milliseconds
98% <= 25,62 milliseconds
99% <= 46,87 milliseconds
99.9% <= 239,61 milliseconds
mean rate = 34,34 calls/second
1-minute rate = 20,98 calls/second
5-minute rate = 15,59 calls/second
15-minute rate = 14,54 calls/second
min = 23,81 milliseconds
max = 325,43 milliseconds
mean = 29,05 milliseconds
stddev = 15,82 milliseconds
median = 24,06 milliseconds
75% <= 25,48 milliseconds
95% <= 60,53 milliseconds
98% <= 72,75 milliseconds
99% <= 91,66 milliseconds
99.9% <= 325,24 milliseconds
```

In the current case, the transpilation has reduced the median evaluation time from 96 millis (for a batch of 1899 data records) to 11 millis, which is almost nine times difference.
In the current case, the transpilation has reduced the median evaluation time from 269 millis (for a batch of 1899 data records) to 24 millis, which is more than eleven times difference.

# License #

Expand Down

0 comments on commit ab3805e

Please sign in to comment.