← AI glossary

Parameter (weight)

The numbers inside a model that are learned during training. “8B” means 8 billion parameters.

A neural network is a collection of billions of numbers called parameters or weights. Training nudges these numbers little by little until the model performs well.

More parameters usually means a more capable model — and more GPU memory and time. B means billion: 8B is 8 billion, 70B is 70 billion.

Example

Llama 3.1 8B has about 8.03 billion parameters; at 2 bytes each (FP16) the weights alone are about 15 GB.

বাংলায়: প্যারামিটার (ওয়েট) — মডেলের ভেতরের সংখ্যা যেগুলো ট্রেনিংয়ে শেখা হয়। “8B” মানে ৮০০ কোটি প্যারামিটার।

Try it on a real AI Computer

JupyterLab opens in about a minute. Pay in taka with bKash, billed by the minute — stop whenever you like.