How to Launch Text Lab
Text Lab runs as an interactive app on the University's High Performance Computing cluster (UBELIX) via the Open OnDemand portal.
Step-by-Step Guide
- Log in to the UBELIX Open OnDemand Portal.
- Navigate to Interactive Apps or Data Science Lab Services and select Text Lab.
- You will see a configuration form. Configure the resource requirements as follows:
Configuration Parameters
| Parameter | Recommended Setting | Description |
|---|---|---|
| Account | gratis |
Selects the type of account to launch Text Lab. |
| SLURM Partition | gpu |
Selects the partition with graphical processing units. |
| QoS (Quality of Service) | job_gpu_preemptable |
Use job_gpu_preemptable for quick tasks with a chance of being disconnected. Use job_gpu if you have a specific allocation. |
| Job Time | 1 to 4 hours |
Estimate how long you need the app. If the timer runs out, the session closes. |
| GPU Type | rtx4090 |
The RTX4090 is powerful and sufficient for most Text Lab tasks (Transcription/OCR). Use A100/H100 only for very large LLM workloads. |
| Number of GPUs | 1 |
One GPU is sufficient for standard usage. |
| WCKey | Optional | Leave blank unless you have a specific project accounting key. |
Advanced Mode
Checking Advanced Mode allows you to specify a "Reservation" if you have code for for a workshop or course. Otherwise, leave this unchecked.
Starting the Session
- Click Launch.
- Wait for the job to start (the status will change from "Queued" to "Running").
- Click Connect to Text Lab to open the interface in your browser.
More info on the configuration parameters
Please check the UBELIX HPC documentation for more info on launching jobs: https://hpc-unibe-ch.github.io/