Inference

Inference is the stage where a trained AI model is actually used to make predictions or decisions.

Key points:

• It happens after training: the model first learns patterns from data, then uses those patterns to answer new inputs.
• Inference can run on many types of hardware (phones, laptops, GPUs, specialized AI chips), but:
  • Bigger models need more powerful hardware to respond quickly.
  • On weaker devices, very large models can be extremely slow compared to running them on cloud servers with high‑end AI accelerators.