Model comparison
Gemma-3-4b-it-Uncensored-DBL-X vs HyperCLOVAX-SEED-Text-Instruct-1.5B
At Q4_K_M, HyperCLOVAX-SEED-Text-Instruct-1.5B is the smaller download — 1,133,974,368 bytes against 2,709,839,808.
From the file· summed bytes, KV per layer
Side by side
| Gemma-3-4b-it-Uncensored-DBL-X | HyperCLOVAX-SEED-Text-Instruct-1.5B | |
|---|---|---|
| Parameters | 4.7B | 1.6B |
| Architecture | gemma3 | llama |
| Layers | 38 | — |
| Native context | 131,072 | — |
| Mixture of experts | no | no |
| Quantizations published | 39 | 1 |
| Smallest quantization | 1.13 GiB | 1.06 GiB |
| Q4_K_M | 2.52 GiB | 1.06 GiB |
| Licence | apache-2.0 | other |
KV cache by context
the term that decides long-context viability
| Context | Gemma-3-4b-it-Uncensored-DBL-X | HyperCLOVAX-SEED-Text-Instruct-1.5B | Ratio |
|---|---|---|---|
| 4,096 | 0.28 GiB | — | — |
| 8,192 | 0.38 GiB | — | — |
| 16,384 | 0.56 GiB | — | — |
| 32,768 | 0.94 GiB | — | — |
| 65,536 | 1.69 GiB | — | — |
| 131,072 | 3.19 GiB | — | — |