Model comparison
rfdetr vs gemma-4-E4B-it-assistant
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| rfdetr | gemma-4-E4B-it-assistant | |
|---|---|---|
| Parameters | 30M | 79M |
| Architecture | rfdetr | gemma4-assistant |
| Layers | — | 4 |
| Native context | — | 131,072 |
| Mixture of experts | no | no |
| Quantizations published | 12 | 9 |
| Smallest quantization | 0.03 GiB | 0.07 GiB |
| Q4_K_M | — | 0.07 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | rfdetr | gemma-4-E4B-it-assistant | Ratio |
|---|---|---|---|
| 4,096 | — | 0.01 GiB | — |
| 8,192 | — | 0.02 GiB | — |
| 16,384 | — | 0.04 GiB | — |
| 32,768 | — | 0.07 GiB | — |
| 65,536 | — | 0.13 GiB | — |
| 131,072 | — | 0.26 GiB | — |