Is embeddinggemma-300m-qat-q4_0-unquantized or embeddinggemma-300m smaller?

Both publish quantizations; compare the table for the exact sizes.

Which uses less memory at long context?

embeddinggemma-300m — its KV cache at 32K context is 1.0x smaller, because of how its attention layers cache.

Model comparison

These two publish different quantization sets; the table below has the exact sizes.

From the file· summed bytes, KV per layer

	embeddinggemma-300m-qat-q4_0-unquantized	embeddinggemma-300m
Parameters	303M	303M
Architecture	gemma-embedding	gemma-embedding
Layers	24	24
Native context	2,048	2,048
Mixture of experts	no	no
Quantizations published	2	10
Smallest quantization	0.21 GiB	0.26 GiB
Q4_K_M	—	—
Licence	gemma	—

the term that decides long-context viability

Context	embeddinggemma-300m-qat-q4_0-unquantized	embeddinggemma-300m	Ratio
4,096	0.04 GiB	0.04 GiB	1.00×
8,192	0.05 GiB	0.05 GiB	1.00×
16,384	0.08 GiB	0.08 GiB	1.00×
32,768	0.14 GiB	0.14 GiB	1.00×
65,536	0.27 GiB	0.27 GiB	1.00×
131,072	0.52 GiB	0.52 GiB	1.00×