-
Notifications
You must be signed in to change notification settings - Fork 18.6k
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
caffe on jetson tk1 #1861
Comments
You could try enabling the DEBUG flag in Makefile.config and see if you get any more useful information. Unfortunately, we have little experience running on this hardware, so we can't provide support here. You may have better luck asking on the caffe-users mailing list. |
A link to a write-up on an updated install: http://jetsonhacks.com/2015/01/17/nvidia-jetson-tk1-caffe-deep-learning-framework/ |
Great THX to Jetsonhaks providing a very nice and easy solution! Thx to erictzeng for advice too! |
@kaaicheung What solved the problem for you? @jetsonhacks do you have any thoughts? I'm building master at 543afd3 and getting the following from
Things seem to be working otherwise; is the problem local to |
@erictzeng I think this issue is due to the Jetson being a 32-bit (ARM) device, and the constant
The solution suggested by Боголюбский Алексей of using 2^29 ( I'm not at all sure that this is the best solution; another value might be better, or some other workaround for 32-bit systems; I'm not sure. Maybe there's a way to add a setting in Fixing this would be really great; small systems like the Jetson are super fun places to run Caffe and it would be nice if it installed easily there. @erictzeng would you consider re-opening this issue and suggesting some preferable solution along the lines of the above? I'll try to reply to the two threads on the email list where this has come up. Thanks @erictzeng @kaaicheung @jetsonhacks! |
Argh; I'm afraid the "fix" was not as successful as I had hoped. While |
@ajschumacher That was a good catch on the 32 bit issue. I guess there will always be a tension between the 64 bit unconstrained implementation versus the 32 bit with more limited resources. |
Thanks @jetsonhacks - credit due to Боголюбский on the mailing list. I also wrote up more about my setup efforts here: http://planspace.org/20150614-the_nvidia_jetson_tk1_with_caffe_on_mnist/ |
To get the LMDB portion of tests to work, make sure to also update examples/mnist/convert_mnist_data.cpp:89:56: warning: large integer implicitly truncated to unsigned type [-Woverflow]
CHECK_EQ(mdb_env_set_mapsize(mdb_env, 1099511627776), MDB_SUCCESS) // 1TB
^ Here is the end of my run: $ make runtest
[...]
[----------] Global test environment tear-down
[==========] 1209 tests from 210 test cases ran. (4425801 ms total)
[ PASSED ] 1209 tests.
YOU HAVE 2 DISABLED TESTS |
I tried to change the LMDB_MAP_SIZE to 536870912 and my second try is 268435456 E1209 10:04:34.680079 4314 convert_imageset.cpp:143] Processed 5475000 files. FYI, I use amazon web service with ubuntu14.04 64bit. I put image data into other partition and I put the lmdb file to another partition. |
Update for my problem, |
Hello, I am having a problem going form lmdb files generated on my mac (64-bit) to the jetson tx1 (32-bit) After a full day of trouble shooting, I've come to the conclusion that lmdb files that work on my TX1 dont work on my mac, and lmdb files that work on my mac don t work on my TX1. I am creating my lmdb files from my mac, and using an external drive to bring them over to the TX1. Does anyone know how I could convert my lmdb file from the 64-bit format to the 32 bit format?? Everything else works smoothly.. for now ^^ Thanks J |
Got an error during runtest
F0213 10:25:10.921640 9387 db.hpp:109] Check failed: mdb_status == 0 (-30792 vs. 0) MDB_MAP_FULL: Environment mapsize limit reached
*** Check failure stack trace: ***
@ 0x44a85060 (unknown)
@ 0x44a84f5c (unknown)
@ 0x44a84b78 (unknown)
@ 0x44a86f98 (unknown)
@ 0x265ec2 caffe::db::LMDBTransaction::Put()
@ 0x21b932 caffe::DBTest_TestWrite_Test<>::TestBody()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
@ 0x230568 testing::internal::HandleExceptionsInMethodIfSupported<>()
Aborted
make: *** [runtest] Error 134
I followed the install instruction mostly from http://petewarden.com/2014/10/25/how-to-run-the-caffe-deep-learning-vision-library-on-nvidias-jetson-mobile-gpu-board/ , except for that I installed some newer version stuff.
Anyone know how to solve this?
Great THX in advance!
The text was updated successfully, but these errors were encountered: