Is gemma-7b or HyperCLOVAX-SEED-Text-Instruct-1.5B smaller?

HyperCLOVAX-SEED-Text-Instruct-1.5B is smaller at Q4_K_M: 1.06 GiB against 4.96 GiB.

Which uses less memory at long context?

We cannot resolve the KV layout for one of these models, so we do not publish a comparison.

Model comparison

At Q4_K_M, HyperCLOVAX-SEED-Text-Instruct-1.5B is the smaller download — 1,133,974,368 bytes against 5,329,758,592.

From the file· summed bytes, KV per layer

	gemma-7b	HyperCLOVAX-SEED-Text-Instruct-1.5B
Parameters	8.5B	1.6B
Architecture	gemma	llama
Layers	28	—
Native context	8,192	—
Mixture of experts	no	no
Quantizations published	31	1
Smallest quantization	2.01 GiB	1.06 GiB
Q4_K_M	4.96 GiB	1.06 GiB
Licence	gemma	other

the term that decides long-context viability

Context	gemma-7b	HyperCLOVAX-SEED-Text-Instruct-1.5B	Ratio
4,096	1.75 GiB	—	—
8,192	3.50 GiB	—	—
16,384	7.00 GiB	—	—
32,768	14.00 GiB	—	—
65,536	28.00 GiB	—	—
131,072	56.00 GiB	—	—