Thank you TG, the problem seemed to be with the helper functions of the dl4j, however i have tried to import the model without the helper function and it imported perfectly and i have received pretty good results. I have saved the serialized model so as to improve resource usage of for running the model. We can either provide this serialized model or one time load the function and save the model. i am trying later approach because saved models take huge amount of memory to store.(around 500mb for 53 mb model). I have yet tested only with VGG16NoTop model (and still testing.) but still there is a problem with helper functions. I will soon resolve the issue and put a pull request for it. This is what i have been working on: https://github.com/asmehra95/tika/blob/master/tika-parsers/src/main/java/org/apache/tika/parser/recognition/dl4j/DL4JImageRecogniser.java
On 21 March 2017 at 22:05, Thamme Gowda (JIRA) <[email protected]> wrote: > > [ https://issues.apache.org/jira/browse/TIKA-2298?page= > com.atlassian.jira.plugin.system.issuetabpanels:comment- > tabpanel&focusedCommentId=15934847#comment-15934847 ] > > Thamme Gowda commented on TIKA-2298: > ------------------------------------ > > [~asmehra95] > Please share a link to your code, I will have a look on this! > > Could you also refer to my example code at https://github.com/ > USCDataScience/dl4j-kerasimport-examples/tree/master/dl4j-import-example > and see what flags to pass to the importer (especially flags to disable > further training)? > > PR to that repo with your VGG16 example would be greatly appreciated! > > > To improve object recognition parser so that it may work without > external RESTful service setup > > ------------------------------------------------------------ > ----------------------------------- > > > > Key: TIKA-2298 > > URL: https://issues.apache.org/jira/browse/TIKA-2298 > > Project: Tika > > Issue Type: Improvement > > Components: parser > > Affects Versions: 1.14 > > Reporter: Avtar Singh > > Labels: ObjectRecognitionParser > > Fix For: 1.15 > > > > Original Estimate: 672h > > Remaining Estimate: 672h > > > > When ObjectRecognitionParser was built to do image recognition, there > wasn't > > good support for Java frameworks. All the popular neural networks were > in > > C++ or python. Since there was nothing that runs within JVM, we tried > > several ways to glue them to Tika (like CLI, JNI, gRPC, REST). > > However, this game is changing slowly now. Deeplearning4j, the most > famous > > neural network library for JVM, now supports importing models that are > > pre-trained in python/C++ based kits [5]. > > *Improvement:* > > It will be nice to have an implementation of ObjectRecogniser that > > doesn't require any external setup(like installation of native libraries > or > > starting REST services). Reasons: easy to distribute and also to cut the > IO > > time. > > > > -- > This message was sent by Atlassian JIRA > (v6.3.15#6346) >
