anton/mlpractical
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Updating wget command and adding links.

Browse Source
This commit is contained in:
Matt Graham 2017-02-13 14:39:43 +00:00
parent 9ea3d06089
commit 6d1dacc265
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
notes/running-jobs-on-the-compute-cluster.md
View File

@ -1,6 +1,6 @@
# Running jobs on the compute cluster
There is a compute cluster now available for submitting jobs to for your MLP assignments. Jobs on the cluster are scheduled using an open source variant of [Oracle Grid Engine](https://en.wikipedia.org/wiki/Oracle_Grid_Engine) which manages allocation of resources to jobs.
There is a [compute cluster](http://computing.help.inf.ed.ac.uk/msc-teaching-cluster) now available for submitting jobs to for your MLP assignments. Jobs on the cluster are scheduled using an open source variant of [Oracle Grid Engine](https://en.wikipedia.org/wiki/Oracle_Grid_Engine) which manages allocation of resources to jobs.
## Accessing the cluster
@ -87,19 +87,19 @@ The script writes all outputted files (model checkpoints, train and validation s
The complete series of commands you would need to run in a DICE terminal to submit the example script as a job on the compute cluster are as follows:
1. Log in to the head node by running
1. Log in to the head node by running
```
ssh [username]@msccluster
```
2. Download the script file from Github to your cluster homespace
2. Download the script file from Github to your cluster homespace
```
wget
wget https://raw.githubusercontent.com/CSTR-Edinburgh/mlpractical/mlp2016-7/master/scripts/example-tf-mnist-train-job.py
```
3. Create an `experiments` directory
3. Create an `experiments` directory
```
mkdir experiments
```
4. Submit the job to the cluster
4. Submit the job to the cluster
```
qsub -v MLP_DATA_DIR='/disk/scratch/mlp/data',OUTPUT_DIR='$HOME/experiments' example-tf-mnist-train-job.py
```