New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[TIKA-DL] Image recognition powered by deeplearning4j and InceptionV3 #165

Merged
merged 13 commits into from May 9, 2017

Conversation

Projects
None yet
6 participants
@thammegowda
Contributor

thammegowda commented Apr 3, 2017

Summary

  • added tika-dl module which depends on deeplearning4j library. This module will produce an addon with all the DL4J dependencies and its native dependencies which may be optionally added to classpath by user to make use of it
    • By default, the build system includes native libs for all major platforms (such as Linux, Windows, OSX, Android/ARM)
    • Unnecessary native libs can be easily excluded by setting the target platform as -Djavacpp.platform=<target> during the build
    • Permissible target values = {android-arm, linux-x86_64, macosx-x86_64, windows-x86_64, etc.}
  • added DL4JInceptionV3Net.java which provides Image recognition features using InceptionV3.
    • Similar to VGG-16 model in #159, VGG-16 model is huuuuge (over 500MB to download)and requires plenty of RAM (~3GB) to run. The beauty of Inception-V3 model is that it is just 90MB to download and requires ~400MB to run
    • No setup required. This implementation is configured to download the model when it runs the first time. It downloads from our USCDataScience's repo
    • It is flexible. Offers plenty of settings to change them. Look for @Field annotation in the code
  • added a Test case to test the above implementation

How to Test

  1. Build the code : mvn package or mvn package -DskipTests or mvn package -DskipTests -Djavacpp.platform=<>
  2. Run:
java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar \
 org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml dog.jpg

Note:

Tested on macosx-x86_64 platform, we have to test on linux-x86_64 and windows-x86_64 before it gets merged.
Feedback/Critiques are welcome.

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann Apr 3, 2017

Contributor

wow great work @thammegowda going to review this now

Contributor

chrismattmann commented Apr 3, 2017

wow great work @thammegowda going to review this now

@KranthiGV

This comment has been minimized.

Show comment
Hide comment
@KranthiGV

KranthiGV Apr 3, 2017

Contributor

Looks interesting! Will test it on linux x86_64 by tonight.

Contributor

KranthiGV commented Apr 3, 2017

Looks interesting! Will test it on linux x86_64 by tonight.

@KranthiGV

This comment has been minimized.

Show comment
Hide comment
@KranthiGV

KranthiGV Apr 3, 2017

Contributor

Testing on: Ubuntu 14.04 LTS
I have compiled successfully but nevertheless not able to run the example. Can someone help me? Is it a mistake of mine?

timberners@galileo:~/Desktop/gsoc/issues/tika$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar org.apache.tika.cli.TikaCLI --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml dog.jpg
WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET https://media.githubusercontent.com/media/USCDataScience/dl4j-kerasimport-examples/master/dl4j-import-example/data/inception-model-weights.h5
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache exist at /home/timberners/Desktop/gsoc/issues/tika/.tmp-inception/inceptionv3-model.json. Not downloading it
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 127 ms to scan 1 urls, producing 29 keys and 189 values 
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Linux]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fc279b678e4, pid=10027, tid=0x00007fc27b5fa700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_121-b13) (build 1.8.0_121-b13)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.121-b13 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# An error report file with more information is saved as:
# /home/timberners/Desktop/gsoc/issues/tika/hs_err_pid10027.log
*** Error in `java': malloc(): memory corruption: 0x00007fc2d4d7dd60 ***
Aborted (core dumped)

The contents of hs_err_pid10027.log are:

#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fc279b678e4, pid=10027, tid=0x00007fc27b5fa700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_121-b13) (build 1.8.0_121-b13)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.121-b13 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#

---------------  T H R E A D  ---------------

Current thread (0x00007fc2d420b800):  JavaThread "JavaCPP Deallocator" daemon [_thread_in_native, id=10145, stack(0x00007fc27b4fa000,0x00007fc27b5fb000)]

siginfo: si_signo: 11 (SIGSEGV), si_code: 128 (SI_KERNEL), si_addr: 0x0000000000000000
((Truncated))
Contributor

KranthiGV commented Apr 3, 2017

Testing on: Ubuntu 14.04 LTS
I have compiled successfully but nevertheless not able to run the example. Can someone help me? Is it a mistake of mine?

timberners@galileo:~/Desktop/gsoc/issues/tika$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar org.apache.tika.cli.TikaCLI --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml dog.jpg
WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET https://media.githubusercontent.com/media/USCDataScience/dl4j-kerasimport-examples/master/dl4j-import-example/data/inception-model-weights.h5
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache exist at /home/timberners/Desktop/gsoc/issues/tika/.tmp-inception/inceptionv3-model.json. Not downloading it
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 127 ms to scan 1 urls, producing 29 keys and 189 values 
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Linux]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fc279b678e4, pid=10027, tid=0x00007fc27b5fa700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_121-b13) (build 1.8.0_121-b13)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.121-b13 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# An error report file with more information is saved as:
# /home/timberners/Desktop/gsoc/issues/tika/hs_err_pid10027.log
*** Error in `java': malloc(): memory corruption: 0x00007fc2d4d7dd60 ***
Aborted (core dumped)

The contents of hs_err_pid10027.log are:

#
# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fc279b678e4, pid=10027, tid=0x00007fc27b5fa700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_121-b13) (build 1.8.0_121-b13)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.121-b13 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#

---------------  T H R E A D  ---------------

Current thread (0x00007fc2d420b800):  JavaThread "JavaCPP Deallocator" daemon [_thread_in_native, id=10145, stack(0x00007fc27b4fa000,0x00007fc27b5fb000)]

siginfo: si_signo: 11 (SIGSEGV), si_code: 128 (SI_KERNEL), si_addr: 0x0000000000000000
((Truncated))
@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 3, 2017

Contributor

@KranthiGV
Hmm, native C code in libhdf5.so

# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Failed to write core dump. Core dumps have been disabled. 
To enable core dumping, try "ulimit -c unlimited" before starting Java again

Could you run ulimit -c unlimited and try again.

Contributor

thammegowda commented Apr 3, 2017

@KranthiGV
Hmm, native C code in libhdf5.so

# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Failed to write core dump. Core dumps have been disabled. 
To enable core dumping, try "ulimit -c unlimited" before starting Java again

Could you run ulimit -c unlimited and try again.

@saudet

This comment has been minimized.

Show comment
Hide comment
@saudet

saudet Apr 3, 2017

/cc @agibsonccc @turambar Is there any known issues with that example?

saudet commented Apr 3, 2017

/cc @agibsonccc @turambar Is there any known issues with that example?

@turambar

This comment has been minimized.

Show comment
Hide comment
@turambar

turambar Apr 3, 2017

@saudet not that I am aware of.

turambar commented Apr 3, 2017

@saudet not that I am aware of.

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 3, 2017

Contributor

I reran the test on OSX.

Input image

Lion

Log

$ wget https://isafari.nathab.com/assets/wildlife/animals/Mammals-lion-a45dcdd8d516caeb89144bd248cff643.jpg -O lion.jpg

$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar  org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml lion.jpg

WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache exists at /Users/thammegr/work/asf/tika/.tmp-inception/inception-model-weights.h5. Not downloading it
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache exists at /Users/thammegr/work/asf/tika/.tmp-inception/inceptionv3-model.json. Not downloading it
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 128 ms to scan 1 urls, producing 29 keys and 189 values
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Mac OS X]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
INFO  Reflections took 1669 ms to scan 3 urls, producing 3474 keys and 15476 values
WARN  Layer "batchnormalization_1" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_2" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
....
WARN  Layer "batchnormalization_94" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Model cannot be trained: output layer predictions is not an IOutputLayer (no loss function specified)
WARN  Model cannot be trained: output predictions is not an IOutputLayer (no loss function specified)
INFO  Reflections took 109 ms to scan 1 urls, producing 371 keys and 1443 values
INFO  Loaded the Inception model. Time taken=5055ms
INFO  minConfidence = 0.015, topN=10
INFO  Recogniser = org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO  Recogniser Available = true
<?xml version="1.0" encoding="UTF-8"?><html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta name="org.apache.tika.parser.recognition.object.rec.impl" content="org.apache.tika.dl.imagerec.DL4JInceptionV3Net"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.CompositeParser"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.recognition.ObjectRecognitionParser"/>
<meta name="resourceName" content="lion.jpg"/>
<meta name="Content-Length" content="122487"/>
<meta name="OBJECT" content="lion (0.95382)"/>
<meta name="Content-Type" content="image/jpeg"/>
<title/>
</head>
<body><ol id="objects"><li id="291"> lion [en](confidence = 0.953824 )</li>
</ol>
</body></html

Yes, it found lion

<ol id="objects"><li id="291"> lion [en](confidence = 0.953824 )</li>
</ol>
Contributor

thammegowda commented Apr 3, 2017

I reran the test on OSX.

Input image

Lion

Log

$ wget https://isafari.nathab.com/assets/wildlife/animals/Mammals-lion-a45dcdd8d516caeb89144bd248cff643.jpg -O lion.jpg

$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar  org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml lion.jpg

WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache exists at /Users/thammegr/work/asf/tika/.tmp-inception/inception-model-weights.h5. Not downloading it
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache exists at /Users/thammegr/work/asf/tika/.tmp-inception/inceptionv3-model.json. Not downloading it
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 128 ms to scan 1 urls, producing 29 keys and 189 values
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Mac OS X]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
INFO  Reflections took 1669 ms to scan 3 urls, producing 3474 keys and 15476 values
WARN  Layer "batchnormalization_1" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_2" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
....
WARN  Layer "batchnormalization_94" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Model cannot be trained: output layer predictions is not an IOutputLayer (no loss function specified)
WARN  Model cannot be trained: output predictions is not an IOutputLayer (no loss function specified)
INFO  Reflections took 109 ms to scan 1 urls, producing 371 keys and 1443 values
INFO  Loaded the Inception model. Time taken=5055ms
INFO  minConfidence = 0.015, topN=10
INFO  Recogniser = org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO  Recogniser Available = true
<?xml version="1.0" encoding="UTF-8"?><html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta name="org.apache.tika.parser.recognition.object.rec.impl" content="org.apache.tika.dl.imagerec.DL4JInceptionV3Net"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.CompositeParser"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.recognition.ObjectRecognitionParser"/>
<meta name="resourceName" content="lion.jpg"/>
<meta name="Content-Length" content="122487"/>
<meta name="OBJECT" content="lion (0.95382)"/>
<meta name="Content-Type" content="image/jpeg"/>
<title/>
</head>
<body><ol id="objects"><li id="291"> lion [en](confidence = 0.953824 )</li>
</ol>
</body></html

Yes, it found lion

<ol id="objects"><li id="291"> lion [en](confidence = 0.953824 )</li>
</ol>
@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 4, 2017

Contributor

I tested this on my linux box. It appears to me that there is a stability related bug in libhdf5.so and hence the importer functionality is affected.
I ran the above-described test 100 times and found that DL4J crashed 6 times for the following reason:

# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fddefb988e4, pid=15318, tid=0x00007fddf410d700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_101-b13) (build 1.8.0_101-b13)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.101-b13 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Core dump written. Default location: /home/tg/work/projects/apache/tika/core or core.15318
#
# An error report file with more information is saved as:
# /home/tg/work/projects/apache/tika/hs_err_pid15318.log
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.

I tested on Linux x86_64 kernel 4.4.0-21-generic

$ uname -a
Linux hackb0x 4.4.0-21-generic #37-Ubuntu SMP Mon Apr 18 18:33:37 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux

CC : @saudet @turambar @agibsonccc We are glad you are all watching this, please help.

Contributor

thammegowda commented Apr 4, 2017

I tested this on my linux box. It appears to me that there is a stability related bug in libhdf5.so and hence the importer functionality is affected.
I ran the above-described test 100 times and found that DL4J crashed 6 times for the following reason:

# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007fddefb988e4, pid=15318, tid=0x00007fddf410d700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_101-b13) (build 1.8.0_101-b13)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.101-b13 mixed mode linux-amd64 compressed oops)
# Problematic frame:
# C  [libhdf5.so.100+0x1138e4]  H5FL_reg_free+0x84
#
# Core dump written. Default location: /home/tg/work/projects/apache/tika/core or core.15318
#
# An error report file with more information is saved as:
# /home/tg/work/projects/apache/tika/hs_err_pid15318.log
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.java.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.

I tested on Linux x86_64 kernel 4.4.0-21-generic

$ uname -a
Linux hackb0x 4.4.0-21-generic #37-Ubuntu SMP Mon Apr 18 18:33:37 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux

CC : @saudet @turambar @agibsonccc We are glad you are all watching this, please help.

@saudet

This comment has been minimized.

Show comment
Hide comment
@saudet

saudet Apr 4, 2017

saudet commented Apr 4, 2017

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 4, 2017

Contributor

@saudet Thanks for the suggestion! I reran 100 tests and it didn't crash a single time! As you expected I had to increase memory (-Xmx1200m instead of -Xmx400m)

seq 1 100 | while read i; do
     echo "==$i==" >> out;
     java -Dorg.bytedeco.javacpp.nopointergc=true -Xmx1200m \
     -cp .:tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar  \
     org.apache.tika.cli.TikaCLI \
     --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml -m \
     lion.jpg >> out 2>&1 ; done

What would be the fix to reduce memory back to 400MB?

Contributor

thammegowda commented Apr 4, 2017

@saudet Thanks for the suggestion! I reran 100 tests and it didn't crash a single time! As you expected I had to increase memory (-Xmx1200m instead of -Xmx400m)

seq 1 100 | while read i; do
     echo "==$i==" >> out;
     java -Dorg.bytedeco.javacpp.nopointergc=true -Xmx1200m \
     -cp .:tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar  \
     org.apache.tika.cli.TikaCLI \
     --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml -m \
     lion.jpg >> out 2>&1 ; done

What would be the fix to reduce memory back to 400MB?

@KranthiGV

This comment has been minimized.

Show comment
Hide comment
@KranthiGV

KranthiGV Apr 4, 2017

Contributor

It worked fine when I changed the file size limits (ulimit -f unlimited).
Twice (out of 50 times), it threw a pointer related exception which was fixed when I set org.bytedeco.javacpp.nopointergc and increased memory.

Contributor

KranthiGV commented Apr 4, 2017

It worked fine when I changed the file size limits (ulimit -f unlimited).
Twice (out of 50 times), it threw a pointer related exception which was fixed when I set org.bytedeco.javacpp.nopointergc and increased memory.

@saudet

This comment has been minimized.

Show comment
Hide comment
@saudet

saudet Apr 4, 2017

@thammegowda We'll have to take a closer look at that.
Could you post this info as an issue at https://github.com/deeplearning4j/deeplearning4j ?
Thanks for your help!!

saudet commented Apr 4, 2017

@thammegowda We'll have to take a closer look at that.
Could you post this info as an issue at https://github.com/deeplearning4j/deeplearning4j ?
Thanks for your help!!

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 9, 2017

Contributor

Update: The issue is fixed in deeplearning4j and it is stable. However, we have to wait till DL4J version 0.8.1 jars become available in maven central.

Contributor

thammegowda commented Apr 9, 2017

Update: The issue is fixed in deeplearning4j and it is stable. However, we have to wait till DL4J version 0.8.1 jars become available in maven central.

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann Apr 9, 2017

Contributor

@thammegowda looks like since deeplearning4j/deeplearning4j#3190 is closed this is testable. I'll be testing this today. Thanks.

Contributor

chrismattmann commented Apr 9, 2017

@thammegowda looks like since deeplearning4j/deeplearning4j#3190 is closed this is testable. I'll be testing this today. Thanks.

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann Apr 9, 2017

Contributor

@thammegowda yes important for 0.8.1 jars to be in Central before committing. Testable now though. The instructions should be updated (can you do so above, and start a wiki page on the Tika wiki please) to git clone and mvn install the 0.8.1 DL4J for now.

Contributor

chrismattmann commented Apr 9, 2017

@thammegowda yes important for 0.8.1 jars to be in Central before committing. Testable now though. The instructions should be updated (can you do so above, and start a wiki page on the Tika wiki please) to git clone and mvn install the 0.8.1 DL4J for now.

@agibsonccc

This comment has been minimized.

Show comment
Hide comment
@agibsonccc

agibsonccc Apr 9, 2017

agibsonccc commented Apr 9, 2017

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 10, 2017

Contributor

@chrismattmann

The instructions should be updated (can you do so above, and start a wiki page on the Tika wiki please) to git clone and mvn install the 0.8.1 DL4J for now.

We could do that but I dont think that is necessary. We should just hang tight till DL4J folks make a public release of version 0.8.1. Reason: to build DL4J (and hence ND4J with matching version) from the source we have to install a bunch of libs, GPU frameworks, and compilers. Should Tika users be bothered about all those details? I dont think so. IMHO we just want to pull from maven repo and use! Do you agree with it Prof.?

Contributor

thammegowda commented Apr 10, 2017

@chrismattmann

The instructions should be updated (can you do so above, and start a wiki page on the Tika wiki please) to git clone and mvn install the 0.8.1 DL4J for now.

We could do that but I dont think that is necessary. We should just hang tight till DL4J folks make a public release of version 0.8.1. Reason: to build DL4J (and hence ND4J with matching version) from the source we have to install a bunch of libs, GPU frameworks, and compilers. Should Tika users be bothered about all those details? I dont think so. IMHO we just want to pull from maven repo and use! Do you agree with it Prof.?

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 10, 2017

Contributor

@agibsonccc
Thanks for the comment. Looks like snapshots are on Sonatype https://oss.sonatype.org/content/repositories/snapshots/org/deeplearning4j/

Can we use them?

Contributor

thammegowda commented Apr 10, 2017

@agibsonccc
Thanks for the comment. Looks like snapshots are on Sonatype https://oss.sonatype.org/content/repositories/snapshots/org/deeplearning4j/

Can we use them?

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda Apr 10, 2017

Contributor

@chrismattmann Now we dont have to do mvn install for DL4J, I just enabled Sonatype snapshots to make it easy testing with 0.8.1-SNAPSHOT version.

I will work on wiki page to document this feature!

Contributor

thammegowda commented Apr 10, 2017

@chrismattmann Now we dont have to do mvn install for DL4J, I just enabled Sonatype snapshots to make it easy testing with 0.8.1-SNAPSHOT version.

I will work on wiki page to document this feature!

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 3, 2017

Contributor

@thammegowda i'm trying this now. I cd into tika-dl and here's what I get:

Downloaded: https://repo.maven.apache.org/maven2/org/bytedeco/javacpp-presets/openblas/0.2.19-1.3/openblas-0.2.19-1.3-macosx-x86_64.jar (12117 KB at 691.6 KB/sec)
Downloaded: https://repo.maven.apache.org/maven2/org/bytedeco/javacpp-presets/openblas/0.2.19-1.3/openblas-0.2.19-1.3-linux-x86_64.jar (11283 KB at 581.6 KB/sec)
Downloading: http://repository.apache.org/snapshots/org/nd4j/nd4j-native/0.8.1-SNAPSHOT/nd4j-native-0.8.1-SNAPSHOT-macosx-x86_64.jar
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 58.779 s
[INFO] Finished at: 2017-05-03T16:34:27-07:00
[INFO] Final Memory: 15M/245M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project tika-dl: Could not resolve dependencies for project org.apache.tika:tika-dl:jar:1.15-SNAPSHOT: Could not find artifact org.nd4j:nd4j-native:jar:macosx-x86_64:0.8.1-SNAPSHOT in oss-sonatype (https://oss.sonatype.org/content/repositories/snapshots/) -> [Help 1]
[ERROR] 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR] 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/DependencyResolutionException
LMC-053601:tika-dl mattmann$ mvn install
[INFO] Scanning for projects...
[INFO]                                                                         
[INFO] ------------------------------------------------------------------------
[INFO] Building tika-dl 1.15-SNAPSHOT
[INFO] ------------------------------------------------------------------------
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 1.656 s
[INFO] Finished at: 2017-05-03T16:35:13-07:00
[INFO] Final Memory: 13M/310M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project tika-dl: Could not resolve dependencies for project org.apache.tika:tika-dl:jar:1.15-SNAPSHOT: Failure to find org.nd4j:nd4j-native:jar:macosx-x86_64:0.8.1-SNAPSHOT in https://oss.sonatype.org/content/repositories/snapshots/ was cached in the local repository, resolution will not be reattempted until the update interval of oss-sonatype has elapsed or updates are forced -> [Help 1]
[ERROR] 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR] 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/DependencyResolutionException
LMC-053601:tika-dl mattmann$ 

any ideas?

Contributor

chrismattmann commented May 3, 2017

@thammegowda i'm trying this now. I cd into tika-dl and here's what I get:

Downloaded: https://repo.maven.apache.org/maven2/org/bytedeco/javacpp-presets/openblas/0.2.19-1.3/openblas-0.2.19-1.3-macosx-x86_64.jar (12117 KB at 691.6 KB/sec)
Downloaded: https://repo.maven.apache.org/maven2/org/bytedeco/javacpp-presets/openblas/0.2.19-1.3/openblas-0.2.19-1.3-linux-x86_64.jar (11283 KB at 581.6 KB/sec)
Downloading: http://repository.apache.org/snapshots/org/nd4j/nd4j-native/0.8.1-SNAPSHOT/nd4j-native-0.8.1-SNAPSHOT-macosx-x86_64.jar
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 58.779 s
[INFO] Finished at: 2017-05-03T16:34:27-07:00
[INFO] Final Memory: 15M/245M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project tika-dl: Could not resolve dependencies for project org.apache.tika:tika-dl:jar:1.15-SNAPSHOT: Could not find artifact org.nd4j:nd4j-native:jar:macosx-x86_64:0.8.1-SNAPSHOT in oss-sonatype (https://oss.sonatype.org/content/repositories/snapshots/) -> [Help 1]
[ERROR] 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR] 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/DependencyResolutionException
LMC-053601:tika-dl mattmann$ mvn install
[INFO] Scanning for projects...
[INFO]                                                                         
[INFO] ------------------------------------------------------------------------
[INFO] Building tika-dl 1.15-SNAPSHOT
[INFO] ------------------------------------------------------------------------
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 1.656 s
[INFO] Finished at: 2017-05-03T16:35:13-07:00
[INFO] Final Memory: 13M/310M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal on project tika-dl: Could not resolve dependencies for project org.apache.tika:tika-dl:jar:1.15-SNAPSHOT: Failure to find org.nd4j:nd4j-native:jar:macosx-x86_64:0.8.1-SNAPSHOT in https://oss.sonatype.org/content/repositories/snapshots/ was cached in the local repository, resolution will not be reattempted until the update interval of oss-sonatype has elapsed or updates are forced -> [Help 1]
[ERROR] 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR] 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/DependencyResolutionException
LMC-053601:tika-dl mattmann$ 

any ideas?

@agibsonccc

This comment has been minimized.

Show comment
Hide comment
@agibsonccc

agibsonccc May 3, 2017

Our mac snapshots are broken /cc @huitseeker working on it . For now just use nd4j-native if you're building snapshots.

agibsonccc commented May 3, 2017

Our mac snapshots are broken /cc @huitseeker working on it . For now just use nd4j-native if you're building snapshots.

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 4, 2017

Contributor

welp, git pull'ed still broke for me:

LMC-053601:tika-gh mattmann$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar  org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml /Users/mattmann/Pictures/0718090901.jpg
WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET https://raw.githubusercontent.com/USCDataScience/dl4j-kerasimport-examples/98ec48b56a5b8fb7d54a2994ce9cb23bfefac821/dl4j-import-example/data/inception-model-weights.h5
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET jar:file:/Users/mattmann/git/tika-gh/tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar!/org/apache/tika/dl/imagerec/inceptionv3-model.json
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 103 ms to scan 1 urls, producing 29 keys and 192 values 
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Mac OS X]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DAudio.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libAppleScriptEngine.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3D.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/usr/lib/java/libjdns_sd.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libmlib_jai.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DUtils.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
INFO  Reflections took 1439 ms to scan 12 urls, producing 3714 keys and 16629 values 
WARN  Layer "batchnormalization_1" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_2" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_3" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_4" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_5" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_9" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_7" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_10" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_6" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_8" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_11" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_12" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_16" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_14" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_17" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_13" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_15" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_18" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_19" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_23" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_21" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_24" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_20" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_22" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_25" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_26" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_28" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_29" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_27" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_30" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_35" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_36" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_32" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_37" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_33" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_38" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_31" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_34" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_39" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_40" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_45" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_46" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_42" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_47" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_43" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_48" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_41" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_44" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_49" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_50" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_55" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_56" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_52" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_57" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_53" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_58" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_51" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_54" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_59" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_60" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_65" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_66" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_62" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_67" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_63" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_68" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_61" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_64" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_69" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_70" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_73" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_74" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_71" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_75" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_72" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_76" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_81" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_78" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_82" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_79" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_80" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_83" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_84" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_77" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_85" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_90" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_87" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_91" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_88" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_89" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_92" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_93" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_86" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_94" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Model cannot be trained: output layer predictions is not an IOutputLayer (no loss function specified)
WARN  Model cannot be trained: output predictions is not an IOutputLayer (no loss function specified)
INFO  Reflections took 104 ms to scan 1 urls, producing 375 keys and 1457 values 
INFO  Loaded the Inception model. Time taken=3723ms
INFO  minConfidence = 0.015, topN=10
INFO  Recogniser = org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO  Recogniser Available = true
WARN  NO objects
LMC-053601:tika-gh mattmann$ 
Contributor

chrismattmann commented May 4, 2017

welp, git pull'ed still broke for me:

LMC-053601:tika-gh mattmann$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar  org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml /Users/mattmann/Pictures/0718090901.jpg
WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET https://raw.githubusercontent.com/USCDataScience/dl4j-kerasimport-examples/98ec48b56a5b8fb7d54a2994ce9cb23bfefac821/dl4j-import-example/data/inception-model-weights.h5
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET jar:file:/Users/mattmann/git/tika-gh/tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar!/org/apache/tika/dl/imagerec/inceptionv3-model.json
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 103 ms to scan 1 urls, producing 29 keys and 192 values 
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Mac OS X]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DAudio.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libAppleScriptEngine.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3D.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/usr/lib/java/libjdns_sd.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libmlib_jai.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DUtils.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:428)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:383)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:373)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:120)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1047)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:578)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:562)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
INFO  Reflections took 1439 ms to scan 12 urls, producing 3714 keys and 16629 values 
WARN  Layer "batchnormalization_1" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_2" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_3" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_4" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_5" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_9" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_7" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_10" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_6" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_8" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_11" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_12" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_16" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_14" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_17" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_13" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_15" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_18" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_19" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_23" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_21" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_24" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_20" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_22" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_25" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_26" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_28" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_29" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_27" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_30" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_35" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_36" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_32" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_37" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_33" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_38" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_31" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_34" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_39" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_40" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_45" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_46" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_42" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_47" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_43" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_48" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_41" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_44" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_49" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_50" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_55" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_56" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_52" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_57" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_53" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_58" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_51" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_54" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_59" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_60" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_65" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_66" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_62" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_67" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_63" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_68" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_61" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_64" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_69" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_70" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_73" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_74" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_71" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_75" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_72" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_76" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_81" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_78" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_82" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_79" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_80" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_83" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_84" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_77" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_85" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_90" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_87" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_91" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_88" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_89" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_92" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_93" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_86" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_94" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Model cannot be trained: output layer predictions is not an IOutputLayer (no loss function specified)
WARN  Model cannot be trained: output predictions is not an IOutputLayer (no loss function specified)
INFO  Reflections took 104 ms to scan 1 urls, producing 375 keys and 1457 values 
INFO  Loaded the Inception model. Time taken=3723ms
INFO  minConfidence = 0.015, topN=10
INFO  Recogniser = org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO  Recogniser Available = true
WARN  NO objects
LMC-053601:tika-gh mattmann$ 
@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 4, 2017

Contributor

So we got this error now

org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DAudio.jnilib]

Contributor

thammegowda commented May 4, 2017

So we got this error now

org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DAudio.jnilib]

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 4, 2017

Contributor

rolling back to 0.8.0 dl4j fixes the issue for me:

LMC-053601:tika-gh mattmann$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar   org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml  https://images-na.ssl-images-amazon.com/images/G/01/img15/pet-products/small-tiles/23695_pets_vertical_store_dogs_small_tile_8._CB312176604_.jpg
WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET https://raw.githubusercontent.com/USCDataScience/dl4j-kerasimport-examples/98ec48b56a5b8fb7d54a2994ce9cb23bfefac821/dl4j-import-example/data/inception-model-weights.h5
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET jar:file:/Users/mattmann/git/tika-gh/tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar!/org/apache/tika/dl/imagerec/inceptionv3-model.json
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 94 ms to scan 1 urls, producing 29 keys and 189 values 
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Mac OS X]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DAudio.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libAppleScriptEngine.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3D.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/usr/lib/java/libjdns_sd.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libmlib_jai.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DUtils.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
INFO  Reflections took 1501 ms to scan 12 urls, producing 3710 keys and 16614 values 
WARN  Layer "batchnormalization_1" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_2" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_3" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_4" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_5" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_9" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_7" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_10" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_6" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_8" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_11" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_12" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_16" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_14" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_17" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_13" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_15" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_18" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_19" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_23" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_21" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_24" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_20" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_22" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_25" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_26" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_28" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_29" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_27" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_30" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_35" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_36" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_32" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_37" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_33" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_38" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_31" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_34" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_39" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_40" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_45" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_46" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_42" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_47" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_43" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_48" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_41" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_44" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_49" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_50" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_55" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_56" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_52" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_57" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_53" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_58" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_51" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_54" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_59" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_60" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_65" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_66" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_62" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_67" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_63" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_68" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_61" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_64" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_69" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_70" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_73" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_74" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_71" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_75" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_72" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_76" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_81" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_78" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_82" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_79" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_80" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_83" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_84" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_77" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_85" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_90" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_87" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_91" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_88" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_89" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_92" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_93" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_86" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_94" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Model cannot be trained: output layer predictions is not an IOutputLayer (no loss function specified)
WARN  Model cannot be trained: output predictions is not an IOutputLayer (no loss function specified)
INFO  Reflections took 107 ms to scan 1 urls, producing 371 keys and 1443 values 
INFO  Loaded the Inception model. Time taken=4944ms
INFO  minConfidence = 0.015, topN=10
INFO  Recogniser = org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO  Recogniser Available = true
<?xml version="1.0" encoding="UTF-8"?><html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta name="org.apache.tika.parser.recognition.object.rec.impl" content="org.apache.tika.dl.imagerec.DL4JInceptionV3Net"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.CompositeParser"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.recognition.ObjectRecognitionParser"/>
<meta name="resourceName" content="23695_pets_vertical_store_dogs_small_tile_8._CB312176604_.jpg"/>
<meta name="Content-Length" content="56996"/>
<meta name="OBJECT" content="beagle (0.60227)"/>
<meta name="OBJECT" content="tennis_ball (0.11380)"/>
<meta name="OBJECT" content="Walker_hound (0.05603)"/>
<meta name="OBJECT" content="Labrador_retriever (0.02694)"/>
<meta name="OBJECT" content="bluetick (0.02557)"/>
<meta name="OBJECT" content="English_foxhound (0.02299)"/>
<meta name="Content-Type" content="image/jpeg"/>
<title/>
</head>
<body><ol id="objects">	<li id="162"> beagle [en](confidence = 0.602268 )</li>
	<li id="852"> tennis_ball [en](confidence = 0.113798 )</li>
	<li id="166"> Walker_hound [en](confidence = 0.056029 )</li>
	<li id="208"> Labrador_retriever [en](confidence = 0.026943 )</li>
	<li id="164"> bluetick [en](confidence = 0.025571 )</li>
	<li id="167"> English_foxhound [en](confidence = 0.022993 )</li>
</ol>
</body></html>LMC-053601:tika-gh mattmann$ 

Contributor

chrismattmann commented May 4, 2017

rolling back to 0.8.0 dl4j fixes the issue for me:

LMC-053601:tika-gh mattmann$ java -Xmx400m -cp ./tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar:tika-app/target/tika-app-1.15-SNAPSHOT.jar   org.apache.tika.cli.TikaCLI  --config=tika-dl/src/test/resources/org/apache/tika/dl/imagerec/dl4j-inception3-config.xml  https://images-na.ssl-images-amazon.com/images/G/01/img15/pet-products/small-tiles/23695_pets_vertical_store_dogs_small_tile_8._CB312176604_.jpg
WARN  JBIG2ImageReader not loaded. jbig2 files will be ignored
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET https://raw.githubusercontent.com/USCDataScience/dl4j-kerasimport-examples/98ec48b56a5b8fb7d54a2994ce9cb23bfefac821/dl4j-import-example/data/inception-model-weights.h5
WARN  File org/apache/tika/dl/imagerec/inceptionv3-model.json not found in local file system. Asking the classloader
INFO  Cache doesn't exist. Going to make a copy
INFO  This might take a while! GET jar:file:/Users/mattmann/git/tika-gh/tika-dl/target/tika-dl-1.15-SNAPSHOT-jar-with-dependencies.jar!/org/apache/tika/dl/imagerec/inceptionv3-model.json
WARN  File org/apache/tika/dl/imagerec/imagenet_incpetionv3_class_index.json not found in local file system. Asking the classloader
INFO  Going to load Inception network...
INFO  Loaded [CpuBackend] backend
INFO  Number of threads used for NativeOps: 4
INFO  Reflections took 94 ms to scan 1 urls, producing 29 keys and 189 values 
INFO  Number of threads used for BLAS: 4
INFO  Backend used: [CPU]; OS: [Mac OS X]
INFO  Cores: [8]; Memory: [0.3GB];
INFO  Blas vendor: [OPENBLAS]
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DAudio.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libAppleScriptEngine.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3D.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/usr/lib/java/libjdns_sd.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libmlib_jai.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
WARN  could not create Vfs.Dir from url. ignoring the exception and continuing
org.reflections.ReflectionsException: could not create Vfs.Dir from url, no matching UrlType was found [file:/System/Library/Java/Extensions/libJ3DUtils.jnilib]
either use fromURL(final URL url, final List<UrlType> urlTypes) or use the static setDefaultURLTypes(final List<UrlType> urlTypes) or addDefaultURLTypes(UrlType urlType) with your specialized UrlType.
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:109)
	at org.reflections.vfs.Vfs.fromURL(Vfs.java:91)
	at org.reflections.Reflections.scan(Reflections.java:237)
	at org.reflections.Reflections.scan(Reflections.java:204)
	at org.reflections.Reflections.<init>(Reflections.java:129)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.registerSubtypes(NeuralNetConfiguration.java:431)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.configureMapper(NeuralNetConfiguration.java:386)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.initMapper(NeuralNetConfiguration.java:376)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration.<clinit>(NeuralNetConfiguration.java:123)
	at org.deeplearning4j.nn.conf.NeuralNetConfiguration$Builder.build(NeuralNetConfiguration.java:1019)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:559)
	at org.deeplearning4j.nn.conf.ComputationGraphConfiguration$GraphBuilder.addLayer(ComputationGraphConfiguration.java:543)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraphConfiguration(KerasModel.java:428)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:482)
	at org.deeplearning4j.nn.modelimport.keras.KerasModel.getComputationGraph(KerasModel.java:471)
	at org.deeplearning4j.nn.modelimport.keras.KerasModelImport.importKerasModelAndWeights(KerasModelImport.java:178)
	at org.apache.tika.dl.imagerec.DL4JInceptionV3Net.initialize(DL4JInceptionV3Net.java:265)
	at org.apache.tika.parser.recognition.ObjectRecognitionParser.initialize(ObjectRecognitionParser.java:100)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOne(TikaConfig.java:603)
	at org.apache.tika.config.TikaConfig$XmlLoader.loadOverall(TikaConfig.java:525)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:185)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:162)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:137)
	at org.apache.tika.config.TikaConfig.<init>(TikaConfig.java:133)
	at org.apache.tika.cli.TikaCLI.configure(TikaCLI.java:675)
	at org.apache.tika.cli.TikaCLI.process(TikaCLI.java:407)
	at org.apache.tika.cli.TikaCLI.main(TikaCLI.java:143)
INFO  Reflections took 1501 ms to scan 12 urls, producing 3710 keys and 16614 values 
WARN  Layer "batchnormalization_1" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_2" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_3" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_4" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_5" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_9" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_7" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_10" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_6" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_8" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_11" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_12" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_16" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_14" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_17" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_13" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_15" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_18" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_19" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_23" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_21" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_24" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_20" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_22" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_25" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_26" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_28" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_29" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_27" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_30" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_35" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_36" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_32" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_37" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_33" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_38" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_31" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_34" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_39" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_40" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_45" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_46" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_42" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_47" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_43" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_48" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_41" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_44" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_49" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_50" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_55" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_56" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_52" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_57" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_53" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_58" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_51" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_54" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_59" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_60" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_65" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_66" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_62" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_67" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_63" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_68" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_61" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_64" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_69" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_70" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_73" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_74" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_71" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_75" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_72" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_76" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_81" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_78" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_82" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_79" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_80" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_83" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_84" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_77" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_85" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_90" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_87" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_91" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_88" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_89" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_92" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_93" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_86" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Layer "batchnormalization_94" momentum has been set but will not be applied unless the updater is set to NESTEROVS.
WARN  Model cannot be trained: output layer predictions is not an IOutputLayer (no loss function specified)
WARN  Model cannot be trained: output predictions is not an IOutputLayer (no loss function specified)
INFO  Reflections took 107 ms to scan 1 urls, producing 371 keys and 1443 values 
INFO  Loaded the Inception model. Time taken=4944ms
INFO  minConfidence = 0.015, topN=10
INFO  Recogniser = org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO  Recogniser Available = true
<?xml version="1.0" encoding="UTF-8"?><html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta name="org.apache.tika.parser.recognition.object.rec.impl" content="org.apache.tika.dl.imagerec.DL4JInceptionV3Net"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.CompositeParser"/>
<meta name="X-Parsed-By" content="org.apache.tika.parser.recognition.ObjectRecognitionParser"/>
<meta name="resourceName" content="23695_pets_vertical_store_dogs_small_tile_8._CB312176604_.jpg"/>
<meta name="Content-Length" content="56996"/>
<meta name="OBJECT" content="beagle (0.60227)"/>
<meta name="OBJECT" content="tennis_ball (0.11380)"/>
<meta name="OBJECT" content="Walker_hound (0.05603)"/>
<meta name="OBJECT" content="Labrador_retriever (0.02694)"/>
<meta name="OBJECT" content="bluetick (0.02557)"/>
<meta name="OBJECT" content="English_foxhound (0.02299)"/>
<meta name="Content-Type" content="image/jpeg"/>
<title/>
</head>
<body><ol id="objects">	<li id="162"> beagle [en](confidence = 0.602268 )</li>
	<li id="852"> tennis_ball [en](confidence = 0.113798 )</li>
	<li id="166"> Walker_hound [en](confidence = 0.056029 )</li>
	<li id="208"> Labrador_retriever [en](confidence = 0.026943 )</li>
	<li id="164"> bluetick [en](confidence = 0.025571 )</li>
	<li id="167"> English_foxhound [en](confidence = 0.022993 )</li>
</ol>
</body></html>LMC-053601:tika-gh mattmann$ 

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 4, 2017

Contributor

After downgrading to 0.8.0, then running mvn test in tika-dl 2x, I got SEGFAULT, however on the 3rd time, I got:

-------------------------------------------------------
 T E S T S
-------------------------------------------------------
Running org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 5.255 sec - in org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest
HDF5: infinite loop closing library
SUREFIRE-859:       L,D_top,S_top,T_top,F,P,P,FD,E,SL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 8.742 s
[INFO] Finished at: 2017-05-03T22:10:02-07:00
[INFO] Final Memory: 31M/964M
[INFO] ------------------------------------------------------------------------
LMC-053601:tika-dl mattmann$ 

I'm not getting a warm fuzzy here that this is stable enough to include in Tika, @thammegowda. Tests don't pass and no objects recognized on my machine using 0.8.1-SNAPSHOT. I was able to (after a lot of heavy lifting), build 0.8.1-SNAPSHOT (nd4j and libnd4j natives) - wasn't too terrible - however each time, on the first run, I would get a SEGFAULT, then after that it never SEGFAULTED, but never produced any recognized objects. With 0.8.0, it recognizes objects, but periodically SEGFAULTS.

Not sure what to do here, I think this code needs major stabilization, not sure where the issue is. We should probably hold off on inclusion in 1.15.

Contributor

chrismattmann commented May 4, 2017

After downgrading to 0.8.0, then running mvn test in tika-dl 2x, I got SEGFAULT, however on the 3rd time, I got:

-------------------------------------------------------
 T E S T S
-------------------------------------------------------
Running org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 5.255 sec - in org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest
HDF5: infinite loop closing library
SUREFIRE-859:       L,D_top,S_top,T_top,F,P,P,FD,E,SL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL

Results :

Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 8.742 s
[INFO] Finished at: 2017-05-03T22:10:02-07:00
[INFO] Final Memory: 31M/964M
[INFO] ------------------------------------------------------------------------
LMC-053601:tika-dl mattmann$ 

I'm not getting a warm fuzzy here that this is stable enough to include in Tika, @thammegowda. Tests don't pass and no objects recognized on my machine using 0.8.1-SNAPSHOT. I was able to (after a lot of heavy lifting), build 0.8.1-SNAPSHOT (nd4j and libnd4j natives) - wasn't too terrible - however each time, on the first run, I would get a SEGFAULT, then after that it never SEGFAULTED, but never produced any recognized objects. With 0.8.0, it recognizes objects, but periodically SEGFAULTS.

Not sure what to do here, I think this code needs major stabilization, not sure where the issue is. We should probably hold off on inclusion in 1.15.

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 4, 2017

Contributor

I vote +1 to hold this PR for the next release (i.e 1.16) and continue releasing 1.15 without it.

Contributor

thammegowda commented May 4, 2017

I vote +1 to hold this PR for the next release (i.e 1.16) and continue releasing 1.15 without it.

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 7, 2017

Contributor

@thammegowda see: deeplearning4j/deeplearning4j#3375 if 0.8.1 DL4J is released soon then we can include tika-dl which I can now get to build regularly.

Contributor

chrismattmann commented May 7, 2017

@thammegowda see: deeplearning4j/deeplearning4j#3375 if 0.8.1 DL4J is released soon then we can include tika-dl which I can now get to build regularly.

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 7, 2017

Contributor

@thammegowda as an FYI here. I see in our https://github.com/USCDataScience/dl4j-kerasimport-examples repo you have linked to several models that we could try and use via Keras import here, including an audio one, and the VGG one e.g., #159. I think we should - if deeplearning4j/deeplearning4j#3375 (comment) works and if we can add support for VGG and the audio model - include tika-dl in 1.15. It would be a HUGE improvement. I think the code we have currently in this PR for Tika-DL though doesn't need to be specific to Inception - it would work the same with the other models - right?

Contributor

chrismattmann commented May 7, 2017

@thammegowda as an FYI here. I see in our https://github.com/USCDataScience/dl4j-kerasimport-examples repo you have linked to several models that we could try and use via Keras import here, including an audio one, and the VGG one e.g., #159. I think we should - if deeplearning4j/deeplearning4j#3375 (comment) works and if we can add support for VGG and the audio model - include tika-dl in 1.15. It would be a HUGE improvement. I think the code we have currently in this PR for Tika-DL though doesn't need to be specific to Inception - it would work the same with the other models - right?

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 7, 2017

Contributor

@chrismattmann Exactly that's the vision for this module - to take full advantage of all those models in all models in https://github.com/fchollet/deep-learning-models/tree/8d8f54a9e1482c3642a55a694c2cde8f26562a06 for audio, video, image, text ....

if we "get DL4J and keras model importer into tika's class path" AND "make it work 💯/💯 on all major platforms like OSX, LINUX and WINDOWS, then we open up huge possibilities of semantic analysis of documents.

I feel it is very difficult to generalize one parser to work for all those models. Each model has its own preprocessing before feeding to the network (we have to apply it on input).
We can easily add parsers later and see the possibilities of generalizing them

Contributor

thammegowda commented May 7, 2017

@chrismattmann Exactly that's the vision for this module - to take full advantage of all those models in all models in https://github.com/fchollet/deep-learning-models/tree/8d8f54a9e1482c3642a55a694c2cde8f26562a06 for audio, video, image, text ....

if we "get DL4J and keras model importer into tika's class path" AND "make it work 💯/💯 on all major platforms like OSX, LINUX and WINDOWS, then we open up huge possibilities of semantic analysis of documents.

I feel it is very difficult to generalize one parser to work for all those models. Each model has its own preprocessing before feeding to the network (we have to apply it on input).
We can easily add parsers later and see the possibilities of generalizing them

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 7, 2017

Contributor

Makes sense to me @thammegowda . I am going to try out the audio one today. if I can get that done and if 0.8.1-1 stabilizes my Mac platform, you can test today too on Linux and if all good, let's commit this to Tika 1.15. We won't include it in the standard build yet, we'll just include it as src, then we can keep growing platforms/models and look for a full feature enabled release and binary release as part of 1.16.

Contributor

chrismattmann commented May 7, 2017

Makes sense to me @thammegowda . I am going to try out the audio one today. if I can get that done and if 0.8.1-1 stabilizes my Mac platform, you can test today too on Linux and if all good, let's commit this to Tika 1.15. We won't include it in the standard build yet, we'll just include it as src, then we can keep growing platforms/models and look for a full feature enabled release and binary release as part of 1.16.

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 7, 2017

Contributor

@chrismattmann

I am going to try out the audio one today.

If you are referring to music_tagger_crnn.py

I am afraid that is not drop-in replacement with the Image recognition parser.
It involves:

  1. export music_tagger_crnn.py weights as .h5 file, model as .json file
  2. create a new parser in tika-dl to deal with audio mime types
  3. and test it

It involves some coding rather than mere testing...! IMHO, it may not be the best use of your time, (we shall let somebody from @USCDataScience raise a new PR, in fact there are lot of students in the community who likes to work on deep learning!)

Contributor

thammegowda commented May 7, 2017

@chrismattmann

I am going to try out the audio one today.

If you are referring to music_tagger_crnn.py

I am afraid that is not drop-in replacement with the Image recognition parser.
It involves:

  1. export music_tagger_crnn.py weights as .h5 file, model as .json file
  2. create a new parser in tika-dl to deal with audio mime types
  3. and test it

It involves some coding rather than mere testing...! IMHO, it may not be the best use of your time, (we shall let somebody from @USCDataScience raise a new PR, in fact there are lot of students in the community who likes to work on deep learning!)

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 7, 2017

Contributor

well let's go then! :) I'll still try and play with it. If 0.8.0-1 passes my tests today, I'm going to commit it, source only. Then follow the plan mentioned above.

Contributor

chrismattmann commented May 7, 2017

well let's go then! :) I'll still try and play with it. If 0.8.0-1 passes my tests today, I'm going to commit it, source only. Then follow the plan mentioned above.

@agibsonccc

This comment has been minimized.

Show comment
Hide comment
@agibsonccc

agibsonccc May 8, 2017

Hey folks - 1 question: How are you guys going to handle the nd4j backend for this? I know it's a minor detail but allowing folks to use the gpu for inference would be pretty neat. Maybe what I'm looking at is a "version 2" of this module? As it is right now nd4j-native-platform will run on every platform (even power and pis!) but for cuda (which some folks will likely end up using) they'll end up kind of limited.

Great job on seeing this through it's already quite a bit of work.

agibsonccc commented May 8, 2017

Hey folks - 1 question: How are you guys going to handle the nd4j backend for this? I know it's a minor detail but allowing folks to use the gpu for inference would be pretty neat. Maybe what I'm looking at is a "version 2" of this module? As it is right now nd4j-native-platform will run on every platform (even power and pis!) but for cuda (which some folks will likely end up using) they'll end up kind of limited.

Great job on seeing this through it's already quite a bit of work.

@saudet

This comment has been minimized.

Show comment
Hide comment
@saudet

saudet May 8, 2017

Users should be able to add the nd4j-cuda-x.x-platform artifact by themselves and it should just work, so it might not be that big of an issue...

saudet commented May 8, 2017

Users should be able to add the nd4j-cuda-x.x-platform artifact by themselves and it should just work, so it might not be that big of an issue...

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 8, 2017

Contributor

Thanks @agibsonccc @saudet I am a thinking to deal with GPUs/Cuda in next version. For now getting CPU backend working stably will do. I like @saudet's suggestion

Contributor

thammegowda commented May 8, 2017

Thanks @agibsonccc @saudet I am a thinking to deal with GPUs/Cuda in next version. For now getting CPU backend working stably will do. I like @saudet's suggestion

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 9, 2017

Contributor

Guys this is now stable enough to commit with 704a039. The DL4J guys released a hotfix 0.8.0-2 with the updates they made to stabilize H5 model weights. I ran it 100x it works fine. Tests pass. I'm going to update tika-bundle and include this as a module in tika-parsers and see if that builds fine. If so, imagine, in 1.15, we could have a DL4J REST server (and use in Tika-Python)! awesome (without an external TensorFlow server - which is great/fine, but this reduces complexity). And in 1.16, we have a module to add further recognizers and parsers. This is great!

Contributor

chrismattmann commented May 9, 2017

Guys this is now stable enough to commit with 704a039. The DL4J guys released a hotfix 0.8.0-2 with the updates they made to stabilize H5 model weights. I ran it 100x it works fine. Tests pass. I'm going to update tika-bundle and include this as a module in tika-parsers and see if that builds fine. If so, imagine, in 1.15, we could have a DL4J REST server (and use in Tika-Python)! awesome (without an external TensorFlow server - which is great/fine, but this reduces complexity). And in 1.16, we have a module to add further recognizers and parsers. This is great!

@chrismattmann chrismattmann self-assigned this May 9, 2017

@chrismattmann chrismattmann added this to the tika-1.15 milestone May 9, 2017

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 9, 2017

Contributor

Awesome.. Happy to see it finally getting into the codebase.

And I will try to get tika+DL4J run on spark / sparkler and our crawling teams can use it for focussed crawling on Hadoop/Spark..

Contributor

thammegowda commented May 9, 2017

Awesome.. Happy to see it finally getting into the codebase.

And I will try to get tika+DL4J run on spark / sparkler and our crawling teams can use it for focussed crawling on Hadoop/Spark..

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 9, 2017

Contributor

OK the last step is now to integrate it into tika-bundle. Wish me luck :/

Contributor

chrismattmann commented May 9, 2017

OK the last step is now to integrate it into tika-bundle. Wish me luck :/

@chrismattmann

This comment has been minimized.

Show comment
Hide comment
@chrismattmann

chrismattmann May 9, 2017

Contributor

Actually I'm going to skip integrating into Tika-Bundle. Not sure exactly how and this can be raised in a future PR. Here comes Tika-DL committing now!

Contributor

chrismattmann commented May 9, 2017

Actually I'm going to skip integrating into Tika-Bundle. Not sure exactly how and this can be raised in a future PR. Here comes Tika-DL committing now!

@chrismattmann chrismattmann merged commit 3207f12 into apache:master May 9, 2017

@agibsonccc

This comment has been minimized.

Show comment
Hide comment
@agibsonccc

agibsonccc May 9, 2017

@thammegowda let us know when you guys need help. For the inference server, I would recommend our new parallelinference server: https://github.com/deeplearning4j/deeplearning4j/blob/master/deeplearning4j-scaleout/deeplearning4j-scaleout-parallelwrapper/src/test/java/org/deeplearning4j/parallelism/ParallelInferenceTest.java#L85 I'll quit pinging you on this PR: More just commenting on your intended roadmap. Let us know if you run in to issues again!

agibsonccc commented May 9, 2017

@thammegowda let us know when you guys need help. For the inference server, I would recommend our new parallelinference server: https://github.com/deeplearning4j/deeplearning4j/blob/master/deeplearning4j-scaleout/deeplearning4j-scaleout-parallelwrapper/src/test/java/org/deeplearning4j/parallelism/ParallelInferenceTest.java#L85 I'll quit pinging you on this PR: More just commenting on your intended roadmap. Let us know if you run in to issues again!

@thammegowda

This comment has been minimized.

Show comment
Hide comment
@thammegowda

thammegowda May 9, 2017

Contributor

@chrismattmann Thanks. I see it is now part of tika-app. I am not sure about bundling stuff either.

Contributor

thammegowda commented May 9, 2017

@chrismattmann Thanks. I see it is now part of tika-app. I am not sure about bundling stuff either.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment