Error running songbird standalone

[swokybio]

Hello,

I am trying to use songbird to analyze the redsea dataset from the tutorial but I am getting an error when calling tensorboard.

The initial step (code below) works well and the output results is generated as expected.

songbird multinomial
--input-biom redsea.biom
--metadata-file redsea_metadata.txt
--formula "Depth+Temperature+Salinity+Oxygen+Fluorescence+Nitrate"
--epochs 10000
--differential-prior 0.5
--training-column Testing
--summary-interval 1 \
--summary-dir results

However, I am having a problem loading tensorboard. I am calling it using tensorboard --logdir . but nothing happens, and I get the following message:

TensorBoard 1.15.0 at http://lewis4-r630-hpc4rc-node361:6006/ (Press CTRL+C to quit) W0713 14:36:23.490031 46913201047296 plugin_event_accumulator.py:294] Found more than one graph event per run, or there was a metagraph containing a graph_def, as well as one or more graph events. Overwriting the graph with the newest event.

Any thoughts on how to use tensorboard to investigate this dataset?

Thank you, Joao

[mortonjt]

That looks like everything is working correctly. What happens when you type in http://lewis4-r630-hpc4rc-node361:6006/ in your browser?

…

On Mon, Jul 13, 2020, 1:56 PM JoaoGabrielMoraes ***@***.***> wrote: Hello, I am trying to use songbird to analyze the redsea dataset from the tutorial but I am getting an error when calling tensorboard. The initial step (code below) works well and the output results is generated as expected. songbird multinomial --input-biom redsea.biom --metadata-file redsea_metadata.txt --formula "Depth+Temperature+Salinity+Oxygen+Fluorescence+Nitrate" --epochs 10000 --differential-prior 0.5 --training-column Testing --summary-interval 1 \ --summary-dir results However, I am having a problem loading tensorboard. I am calling it using tensorboard --logdir . but nothing happens, and I get the following message: TensorBoard 1.15.0 at http://lewis4-r630-hpc4rc-node361:6006/ (Press CTRL+C to quit) W0713 14:36:23.490031 46913201047296 plugin_event_accumulator.py:294] Found more than one graph event per run, or there was a metagraph containing a graph_def, as well as one or more graph events. Overwriting the graph with the newest event. Any thoughts on how to use tensorboard to investigate this dataset? Thank you, Joao — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#133>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXJ5ZATGWIMWPEJ4UF3R3NRFFANCNFSM4OYZ4BKQ> .

[swokybio]

Thanks a lot @mortonjt for your prompt response.

I am running TensorBoard in our University computing cluster and we don't have a browser installed there. Do I need to get one installed?

If I type http://lewis4-r630-hpc4rc-node361:6006/ in the browser of my personal computer I get an error saying that 'Safari can't find the server'.

Thanks, Joao

[mortonjt]

gotcha. Then you'll need to setup an ssh tunnel first While keeping your tensorboard session running on the cluster, try running the following on your local machine ssh -f -N -L 6006:localhost:6006 <your-user-name>@ lewis4-r630-hpc4rc-node361 <http://lewis4-r630-hpc4rc-node361:6006/> That'll hook up tensorboard on the cluster to your local machine, which you can view at localhost:6006 in your browser. P.S. you may need to periodically clean up your ports when running multiple sessions. You can view your processes with `lsof -i tcp:6006` , which you can kill when you are done.

…

On Mon, Jul 13, 2020 at 2:10 PM JoaoGabrielMoraes ***@***.***> wrote: Thanks a lot @mortonjt <https://github.com/mortonjt> for your prompt response. I am running TensorBoard in our University computing cluster and we don't have a browser installed there. Do I need to get one installed? If I type http://lewis4-r630-hpc4rc-node361:6006/ in the browser of my personal computer I get an error saying that 'Safari can't find the server'. Thanks, Joao — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#133 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXP6Y57MSPCZ6DLBEEDR3NSZ7ANCNFSM4OYZ4BKQ> .

[swokybio]

Thanks again @mortonjt!

When I try that I get the following error:

➜ Songbird_example ssh -f -N -L 6006:localhost:6006 jnmpt6@lewis4-r630-hpc4rc-node361 http://lewis4-r630-hpc4rc-node361:6006/\
zsh: no such file or directory: http://lewis4-r630-hpc4rc-node361:6006/>

I think it has something to do with the > sign at the end. Do you know how to not include it?

[fedarko]

It's probably worth noting that there have been security issues with certain versions of TensorFlow (and some of these issues will likely continue, since the 1.x releases will stop being supported as early as October 2020). I don't think (but I can't guarantee at all that) these should be major, but it may be ideal to talk with your HPC staff about how to set up TensorBoard access securely (they may also have done this before, which would make life easier all around :).

[mortonjt]

Agree with @fedarko , getting IT support would certainly help (there are likely some syntax issues with your ssh tunnel).

…

On Mon, Jul 13, 2020 at 2:38 PM Marcus Fedarko ***@***.***> wrote: It's probably worth noting that there have been security issues with certain versions of TensorFlow (and some of these issues will likely continue, since the 1.x releases will stop being supported as early as October 2020 <https://github.com/tensorflow/tensorflow/releases/tag/v1.15.0>). I don't think these should be too major, but it may be ideal to talk with your HPC staff about how to set up TensorBoard access securely (they may also have done this before, which would make life easier all around :). — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#133 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXN4UHCR5FLOVH6VC6DR3NWGFANCNFSM4OYZ4BKQ> .

[swokybio]

Good deal, will do!

Thanks again both of you for your help.

Best, Joao