Model comparison
Llama-Doctor-3.2-3B-Instruct vs Llama-3.1-8B-Instruct
At Q4_K_M, Llama-Doctor-3.2-3B-Instruct is the smaller download — 2,019,377,984 bytes against 4,920,739,168.
From the file· summed bytes, KV per layer
Side by side
| Llama-Doctor-3.2-3B-Instruct | Llama-3.1-8B-Instruct | |
|---|---|---|
| Parameters | 3.2B | 8.0B |
| Architecture | llama | llama |
| Layers | 28 | 32 |
| Native context | 131,072 | 131,072 |
| Mixture of experts | no | no |
| Quantizations published | 48 | 45 |
| Smallest quantization | 0.81 GiB | 2.02 GiB |
| Q4_K_M | 1.88 GiB | 4.58 GiB |
| Licence | mit | llama3.1 |
KV cache by context
the term that decides long-context viability
| Context | Llama-Doctor-3.2-3B-Instruct | Llama-3.1-8B-Instruct | Ratio |
|---|---|---|---|
| 4,096 | 0.44 GiB | 0.50 GiB | 1.14× |
| 8,192 | 0.88 GiB | 1.00 GiB | 1.14× |
| 16,384 | 1.75 GiB | 2.00 GiB | 1.14× |
| 32,768 | 3.50 GiB | 4.00 GiB | 1.14× |
| 65,536 | 7.00 GiB | 8.00 GiB | 1.14× |
| 131,072 | 14.00 GiB | 16.00 GiB | 1.14× |