Tim Allison created TIKA-3503:
---------------------------------
Summary: Figure out how to upgrade dl4j
Key: TIKA-3503
URL: https://issues.apache.org/jira/browse/TIKA-3503
Project: Tika
Issue Type: Task
Reporter: Tim Allison
We're still running with dl4j 1.0.0-beta6. Whenever I try to make general
upgrades, I try out the latest dl4j.
Today, as part of TIKA-3502, I tried to upgrade dl4j to the latest, now
1.0.0-M1.1. I got this exception:
{noformat}
INFO [main] 14:40:27,963 org.apache.tika.dl.imagerec.DL4JInceptionV3Net Cache
exist at
/home/tallison/.tika-dl/models/keras/inception-v3/inception_v3_keras_2.h5. Not
downloading it
INFO [main] 14:40:27,966 org.apache.tika.dl.imagerec.DL4JInceptionV3Net Cache
exist at
/home/tallison/.tika-dl/models/keras/inception-v3/imagenet_class_index.json.
Not downloading it
Warning: Could not load Loader: java.lang.UnsatisfiedLinkError: no jnijavacpp
in java.library.path
INFO [main] 14:40:29,918 org.apache.tika.dl.imagerec.DL4JInceptionV3Net Going
to load Inception network...
WARN [main] 14:40:30,501 org.deeplearning4j.nn.conf.inputs.InputType Assigning
height of 0. Normally this is not valid. Exceptions for this are generally
relatedto model import and unknown dimensions
WARN [main] 14:40:30,501 org.deeplearning4j.nn.conf.inputs.InputType Assigning
width of 0. Normally this is not valid. Exceptions for this are generally
relatedto model import and unknown dimensions
INFO [main] 14:40:30,661 org.nd4j.linalg.factory.Nd4jBackend Loaded
[CpuBackend] backend
INFO [main] 14:40:32,415 org.nd4j.nativeblas.NativeOpsHolder Number of threads
used for linear algebra: 4
INFO [main] 14:40:32,416 org.nd4j.linalg.cpu.nativecpu.CpuNDArrayFactory
Binary level Generic x86 optimization level AVX/AVX2
INFO [main] 14:40:32,420 org.nd4j.nativeblas.Nd4jBlas Number of threads used
for OpenMP BLAS: 4
INFO [main] 14:40:32,489
org.nd4j.linalg.api.ops.executioner.DefaultOpExecutioner Backend used: [CPU];
OS: [Linux]
INFO [main] 14:40:32,489
org.nd4j.linalg.api.ops.executioner.DefaultOpExecutioner Cores: [8]; Memory:
[2.7GB];
INFO [main] 14:40:32,489
org.nd4j.linalg.api.ops.executioner.DefaultOpExecutioner Blas vendor: [OPENBLAS]
INFO [main] 14:40:32,494 org.nd4j.linalg.cpu.nativecpu.CpuBackend Backend
build information:
GCC: "7.5.0"
STD version: 201103L
DEFAULT_ENGINE: samediff::ENGINE_CPU
HAVE_FLATBUFFERS
HAVE_OPENBLAS
INFO [main] 14:40:33,635 org.deeplearning4j.nn.graph.ComputationGraph Starting
ComputationGraph with WorkspaceModes set to [training: ENABLED; inference:
ENABLED], cacheMode set to [NONE]
INFO [main] 14:40:33,897 org.apache.tika.dl.imagerec.DL4JInceptionV3Net Loaded
the Inception model. Time taken=3979ms
INFO [main] 14:40:33,897
org.apache.tika.parser.recognition.ObjectRecognitionParser Recogniser =
org.apache.tika.dl.imagerec.DL4JInceptionV3Net
INFO [main] 14:40:33,897
org.apache.tika.parser.recognition.ObjectRecognitionParser Recogniser Available
= true
[ERROR] Tests run: 1, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 6.785 s
<<< FAILURE! - in org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest
[ERROR] org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest.recognise Time
elapsed: 6.756 s <<< FAILURE!
org.opentest4j.AssertionFailedError: expected: <true> but was: <false>
at
org.apache.tika.dl.imagerec.DL4JInceptionV3NetTest.recognise(DL4JInceptionV3NetTest.java:58)
{noformat}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
