Integrations / Transformers
RunDocumented
Run Transformers models on GPUs
Run a Transformers script as a finite job, or serve a compatible model when you need it to stay available.
The steps follow the tool's documented settings. No tested run is recorded for this page yet.
Why use Badgr for Transformers
- Badgr interface
- Run
- Where Transformers fits
- Model framework
- Why Badgr
- Inference and training
- Use one account for both: a finite job for training or batches, a service when an app needs live answers.
- You pick the shape of the work, and billing follows it: until the job ends, or until you stop the service.
- Your scripts run as they are.
What you can do
- Batch classification or embeddings over a dataset
- Fine-tune a small model with the Trainer
- Serve a compatible model when an app needs live answers
Quickstart: Run a script
badgr run . --cmd "python generate.py" --max-cost 5
How it works
- 1.You run one command
- 2.Badgr checks it and picks a GPU
- 3.Your code runs and logs stream back
- 4.Outputs are saved and billing stops
What you get
- Logs while it runs
- Output files, downloaded with badgr artifacts <id>
- A receipt with cost and duration
Requirements
- The Badgr CLI installed and logged in (npm install -g badgr-cli)
- A funded Badgr balance
- A script and requirements.txt (transformers, torch)
- Hub access for gated models
What it costs
- A job is billed for the time it runs and stops billing when your command exits.
- Pass --max-cost to set the most a job can spend. Badgr stops it if it reaches the cap.
- Badgr shows the rate for the GPU it picked before it starts. This page does not quote a fixed price.
- For live answers, serve a compatible model instead. A service bills until you stop it.
See pricing for how billing works.
Proof
No tested run is recorded for this integration yet. A receipt and example output will be added here after a real run, not before.
Troubleshooting
Job or service?
Use a job when the work ends (training, batches). Use badgr serve when an app must call the model repeatedly.
Out of GPU memory
Load in lower precision, shrink the batch size, or pick a larger GPU.
Slow first start
The model downloads at the start of each job. Large models add minutes before work begins.
Developer reference
Ready to start with Transformers?
Every run or service has a spending limit you set before it starts.