Phriction Projects Wikis Bioimaging And Optics Platform Image Processing Machine Learning Deep Learning SCITAS GPU Clusters Protocols History Version 2 vs 3
Version 2 vs 3
Version 2 vs 3
Content Changes
Content Changes
= First Time Setup =
Create environment
virtualenv --system-site-packages -p python3 env-biop-tf2
== Activate environment ==
source env-biop-tf2/bin/activate
== Install TF ==
cp /tmp/tensorflow-2.3.1-cp37-cp37m-linux_x86_64.whl .
load gcc cuda cudnn python
pip install ~/tensorflow-2.3.1-cp37-cp37m-linux_x86_64.whl
pip install ipython
== Test ==
ipython
import tensorflow as tf
= Run on the Cluster =
Request time (THIS NEEDS TO BE ADDED, as it currently fails, ticket open)
`salloc -t 2:0:0`
Find allocated node
login to the allocated node
Check status of node `htop`
load gcc cuda cudnn python
activate virtualenv
run script
NOTE: Nicola will update on the Jupyter Notebook setup and Oli needs to fix the fact that his account does not have qos=gpu
= First Time Setup =
Create environment
virtualenv --system-site-packages -p python3 env-biop-tf2
== Activate environment ==
source env-biop-tf2/bin/activate
== Install TF ==
```
module load python gcc/8.4.0-cuda mvapich2/2.3.4-cuda py-tensorflow
pip install ipython tensorboard
```
== Test ==
ipython
import tensorflow as tf
= Run on the Cluster =
Request time as ptbiop user
`salloc -t 2:0:0 -A ptbiop`
Find allocated node
login to the allocated node
srun --pty bash
Check status of node `htop`
load gcc cuda cudnn python
activate virtualenv
run script
NOTE: Nicola will update on the Jupyter Notebook setup and Oli needs to fix the fact that his account does not have qos=gpu
= First Time Setup =
Create environment
virtualenv --system-site-packages -p python3 env-biop-tf2
== Activate environment ==
source env-biop-tf2/bin/activate
== Install TF ==
cp /tmp/tensorflow-2.3.1-cp37-cp37m-linux_x86_64.whl .```
module load python gcc /8.4.0-cuda cudnn pythonmvapich2/2.3.4-cuda py-tensorflow
pip install ~/tensorflow-2.3.1-cp37-cp37m-linux_x86_64.whlipython tensorboard
pip install ipython
```
== Test ==
ipython
import tensorflow as tf
= Run on the Cluster =
Request time (THIS NEEDS TO BE ADDED, as it currently fails, ticket open)as ptbiop user
`salloc -t 2:0:0 -A ptbiop`
Find allocated node
login to the allocated node
srun --pty bash
Check status of node `htop`
load gcc cuda cudnn python
activate virtualenv
run script
NOTE: Nicola will update on the Jupyter Notebook setup and Oli needs to fix the fact that his account does not have qos=gpu
c4science · Help