Skip to main content

Jupyterlab

Use these commands to use Juice with JupyterLab on Windows.

Install JupyterLab

pip install jupyterlab 

Visit the Jupyter Labs documentation for more information.

Run JupyterLab with Juice

juice run jupyter lab 

Test Juice JupyterLab Connection

Run this command in a new notebook:

!nvidia-smi 

If successful you should see an output similar to the one below:

Wed Jul 17 11:56:15 2024      
+-----------------------------------------------------------------------------------------+

| NVIDIA-SMI 555.85 Driver Version: 556.12 CUDA Version: 12.5 |

|-----------------------------------------+------------------------+----------------------+

| GPU Name Driver-Model | Bus-Id Disp.A | Volatile Uncorr. ECC |

| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |

| | | MIG M. |

|=========================================+========================+======================|

| 0 NVIDIA GeForce RTX 4090 ERR! | 00000000:01:00.0 Off | Off |

| 0% 45C P0 17W / 450W | 0MiB / 24564MiB | 0% Default |

| | | N/A |

+-----------------------------------------+------------------------+----------------------+



+-----------------------------------------------------------------------------------------+

| Processes: |

| GPU GI CI PID Type Process name GPU Memory |

| ID ID Usage |

|=========================================================================================|

| No running processes found |

+-----------------------------------------------------------------------------------------+

Example (with specific arguments)

To specify GPU ID and number of GPUs:

juice run --gpu-ids [gpu uuid] --num-gpus 1 jupyter lab 
note

These commands cover basic setup and execution. Adjust arguments as needed for your specific environment.