Is this something that will show up in Ollama any time soon to increase context size of local models?

KV quantization has long been available in llama.cpp

	▲	chr15m 4 days ago \| parent [-]
		Yes but the optimisation described has not right?