💨 Vikhr-Gemma-2B-instruct
Мощная инструктивная модель на основе Gemma 2 2B, обученная на русскоязычном датасете GrandMaster-PRO-MAX.
Perplexity (ниже - лучше)
Veles
| Model | Perplexity |
|---|---|
| Q4_K | 4.7254 +/- 0.03867 |
| Q4_0 | 4.8067 +/- 0.03922 |
| Q8_0 | 4.6042 +/- 0.03751 |
| Q4_1 | 4.7798 +/- 0.03933 |
| F32 | 4.6013 +/- 0.03749 |
| Q6_K | 4.6244 +/- 0.03760 |
| BF16 | 4.6015 +/- 0.03749 |
| Q2_K | 5.6819 +/- 0.04737 |
| Q5_0 | 4.6876 +/- 0.03855 |
| Q5_K | 4.6428 +/- 0.03789 |
| Q3_K_S | 5.1485 +/- 0.04257 |
| Q2_K_S | 6.3124 +/- 0.05359 |
| F16 | 4.6013 +/- 0.03749 |
| Q4_K_M | 4.7254 +/- 0.03867 |
| Q5_K_M | 4.6428 +/- 0.03789 |
| Q5_1 | 4.6518 +/- 0.03794 |
| Q4_K_S | 4.7631 +/- 0.03916 |
| Q5_K_S | 4.6509 +/- 0.03803 |
| Q3_K | 4.8339 +/- 0.03965 |
| Q3_K_M | 4.8339 +/- 0.03965 |
| Q3_K_L | 4.7981 +/- 0.03934 |
Wikitext-2
| Model | Perplexity |
|---|---|
| Q4_K | 10.4374 +/- 0.07339 |
| Q4_0 | 10.6480 +/- 0.07452 |
| Q8_0 | 10.1209 +/- 0.07105 |
| Q4_1 | 10.5574 +/- 0.07476 |
| F32 | 10.1191 +/- 0.07099 |
| Q6_K | 10.1503 +/- 0.07117 |
| BF16 | 10.1189 +/- 0.07098 |
| Q2_K | 12.8851 +/- 0.09332 |
| Q5_0 | 10.2551 +/- 0.07251 |
| Q5_K | 10.1975 +/- 0.07184 |
| Q3_K_S | 11.6028 +/- 0.08333 |
| Q2_K_S | 14.7951 +/- 0.10960 |
| F16 | 10.1191 +/- 0.07099 |
| Q4_K_M | 10.4374 +/- 0.07339 |
| Q5_K_M | 10.1975 +/- 0.07184 |
| Q5_1 | 10.2348 +/- 0.07208 |
| Q4_K_S | 10.4924 +/- 0.07386 |
| Q5_K_S | 10.2098 +/- 0.07198 |
| Q3_K | 10.7416 +/- 0.07606 |
| Q3_K_M | 10.7416 +/- 0.07606 |
| Q3_K_L | 10.6242 +/- 0.07506 |