Match model size and quantization to the hardware you actually own
You should be able to demonstrate the result described in the lesson and explain why each step was necessary.
Step-by-step
Inventory the machine
Record total system RAM, available GPU VRAM, operating system, and whether the inference tool can use one or multiple GPUs.
Leave headroom
Do not plan to consume every byte of VRAM or RAM. The operating system, model runtime, context cache, and other applications need memory too.
Start smaller than your maximum
Choose a model that loads comfortably, then measure tokens per second and output quality before moving larger.
Compare quantizations
Try two quantizations of the same model and compare latency, memory use, and your own evaluation prompts.
Save a hardware profile
Keep notes showing which models fit and how they performed so you do not repeat the same experiments.
You should be able to demonstrate the result described in the lesson and explain why each step was necessary.
Troubleshooting
If a command or screen differs from the lesson, stop and verify the version you are using before forcing a change.
What you learned
You now understand the purpose of this part of the system and have a concrete result you can verify before continuing to the next lesson.