Miatz playground

Can You Actually Run This?

Pick a model and your GPU. We do the honest math.

Play it now — free, no account All tools

AIE-110BYOK AICross-Cuttinggen · dev

The Demystify signature

the formula is shown working, not just the verdict — every term in parameter-count times bytes-per-parameter plus KV-cache is visible and adjustable, replacing 'just download and run it' hype with arithmetic anyone can rerun themselves

How it works

What goes in, what comes out

What it does

Learner picks a model's parameter count, a quantization level, and a GPU (dropdown of common consumer/prosumer cards or a custom VRAM entry). The tool runs the real arithmetic live — params times bytes-per-parameter, plus KV-cache-per-token times context length, plus runtime overhead — and renders a fuel-gauge VRAM bar with a pass/fail verdict and an estimated tokens/sec range.

You bring

three clicks — model, quantization, GPU — then drag the context-length slider and watch the bar fill live

You get

a literal fits/doesn't-fit verdict with every number in the formula shown, an estimated tokens/sec range, a suggested quantization if it doesn't fit

You control

model-size selector, quantization toggle (fp16/int8/int4), context-length slider, batch-size slider, GPU picker or custom VRAM input

Where it's used

FREE, top-of-funnel — /playgrounds gallery; feeds directly into AIE-110's graded VRAM-arithmetic exercise

VRAM budgetingquantization fp16/int8/int4KV-cache overheadthroughput vs batch sizecontext-length cost

Optional · watch & try

Go deeper, elsewhere

Hand-picked public explainers and open tools that complement this one — always optional, never required, never graded.

What is LoRA? Low-Rank Adaptation for finetuning LLMs EXPLAINEDExplains the low-rank matrix trick that makes LoRA cheap LoRA for fine-tuning LLMs explained with exampleWorked example applying LoRA to an actual model LLM Fine Tuning Explained in 8 Minutes: LoRA, QLoRA, DPOFast comparison of fine-tuning method families · ~8 min Hugging Face PEFT — LoRA methodsOfficial reference implementation and guide for LoRA/PEFT fine-tuning LLM Model VRAM Calculator (Hugging Face Space)Estimate GPU memory needed to fine-tune or run a given model at a given quantization

Concepts click when you open the machinery.

Three labs are already live and free — the same hands-on style this playground brings to its module.

Play the free labs Take the DSAT