Bring your own training code.
Run pre-training, post-training, and LoRA fine-tuning with your code, data, and resource requirements.
TRAINING PROGRESS
Why Nodus
Focus on the work.
We handle the run.
From finding compatible compute to restoring saved files, Nodus manages the infrastructure around your job.
Compute that fits your code.
Specify your container image, memory, and compute requirements. Nodus finds compatible capacity for the job.
How matching works →Know what the run costs.
Track usage and spending across setup, execution, and recovery, with costs tied to each workload.
Set limits for agent runs →Resume from your checkpoint.
When a run is interrupted, Nodus restores saved files on replacement compute. Your code reloads its checkpoint and continues.
Make training resumable →