Am 07.01.2014 21:28, schrieb Marshall Schor: > On 1/7/2014 12:03 PM, Peter Klügl (JIRA) wrote: >> [ >> https://issues.apache.org/jira/browse/UIMA-2332?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13864394#comment-13864394 >> ] >> >> Peter Klügl commented on UIMA-2332: >> ----------------------------------- >> >> after the latest improvements: >> normal inference: 11.4 times faster >> dynamic anchoring: 13.3 times faster >> >> There are still many possibilities to improve the performance, but I think >> that's enough for now. Maybe I will take another look at it tomorrow and >> then resolve the issue for the next release. > with such good progress, if there is more low-hanging fruit, +1 for you to > "take > a look tomorrow"!
The remaining fruits would require more time (which I currently do not have). Some require new concepts, others only a better implementation of low-level functionality. I will add some more information to the issue and then resolve it. I do not want to optimize the inference for a test script, which does not really resemble realistic ruta scripts (at least not those I create). The test script is just a 1-to-1 translation of some ANNIE NER rules missing much of the stuff ruta has in contrast to jape. If I find the time, then maybe I will profile the inference for some of our rule applications. I asked Philip to rerun a script for the segmentation of clinical discharge letters, and the performance has improved by factor 5. That's not bad, but I think a closer look there will highlight different spots of the inference for optimization. Best, Peter > -Marshall >>> Profile and optimize Ruta inference performance >>> ----------------------------------------------- >>> >>> Key: UIMA-2332 >>> URL: https://issues.apache.org/jira/browse/UIMA-2332 >>> Project: UIMA >>> Issue Type: Improvement >>> Components: ruta >>> Affects Versions: 2.0.0TextMarker >>> Reporter: Peter Klügl >>> Assignee: Peter Klügl >>> Priority: Minor >>> Fix For: 2.1.1ruta >>> >>> >>> Increase the speed of the ruta rule inference. A starting point is the >>> slowdown of UIMA-2330, see RutaTypeMatcher.getMatchingAnnotations() >> >> -- >> This message was sent by Atlassian JIRA >> (v6.1.5#6160) >>
