gemma-4-26b-a4b-heretic-styletune-v2-head

25 of 30 layers cache only a 1,024-token window rather than the full context, on a period of . Figures assume the default configuration; --swa-full disables the saving entirely.

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 13.52 GiB. The real file is 16.03 GiB, because a quantization is a mixture and some tensors are always kept at higher precision. The larger discrepancy is the cache: a flat formula gives 7.50 GiB at 32K context where the real figure is 1.54 GiB, because most of this model's layers cache a fixed window rather than the whole context.

How much VRAM does gemma-4-26b-a4b-heretic-styletune-v2-head need?

Q4_K_M is exactly 17,211,251,808 bytes (16.03 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.

How large is gemma-4-26b-a4b-heretic-styletune-v2-head's KV cache?

1.54 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.

Is gemma-4-26b-a4b-heretic-styletune-v2-head a mixture-of-experts model?

Yes — 128 experts, null routed per token. Every expert must be resident, but only the routed ones are read per token, which is why its memory requirement and its speed behave very differently.

Which quantization of gemma-4-26b-a4b-heretic-styletune-v2-head should I use?

Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.