Model comparison
Qwen2.5-3B-Instruct-Uncensored 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 1,929,903,008.
From the file· summed bytes, KV per layer
Side by side
| Qwen2.5-3B-Instruct-Uncensored | HyperCLOVAX-SEED-Text-Instruct-1.5B | |
|---|---|---|
| Parameters | 3.1B | 1.6B |
| Architecture | qwen2 | llama |
| Layers | 36 | — |
| Native context | 32,768 | — |
| Mixture of experts | no | no |
| Quantizations published | 36 | 1 |
| Smallest quantization | 0.74 GiB | 1.06 GiB |
| Q4_K_M | 1.80 GiB | 1.06 GiB |
| Licence | — | other |
KV cache by context
the term that decides long-context viability
| Context | Qwen2.5-3B-Instruct-Uncensored | HyperCLOVAX-SEED-Text-Instruct-1.5B | Ratio |
|---|---|---|---|
| 4,096 | 0.14 GiB | — | — |
| 8,192 | 0.28 GiB | — | — |
| 16,384 | 0.56 GiB | — | — |
| 32,768 | 1.13 GiB | — | — |
| 65,536 | 2.25 GiB | — | — |
| 131,072 | 4.50 GiB | — | — |