Update README.md
Browse files
README.md
CHANGED
|
@@ -24,6 +24,16 @@ without relying on the cloud.
|
|
| 24 |
|
| 25 |
For the MMLU evaluation, we use a 0-shot CoT setting.
|
| 26 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
## Model Card
|
| 28 |
|
| 29 |
<div align="center">
|
|
|
|
| 24 |
|
| 25 |
For the MMLU evaluation, we use a 0-shot CoT setting.
|
| 26 |
|
| 27 |
+
## Speed
|
| 28 |
+
| Model | Memory(GiB) | i9 14900 | 1+13 8ge4 | rk3588 (16G) | Raspberry PI 5 |
|
| 29 |
+
|--------------------------------------|---------------------|----------|-----------|--------------|----------------|
|
| 30 |
+
| SmallThinker 21B+sparse | 11.47 | 30.19 | 23.03 | 10.84 | 6.61 |
|
| 31 |
+
| SmallThinker 21B+sparse +limited memory | limit 8G | 20.30 | 15.50 | 8.56 | - |
|
| 32 |
+
| Qwen3 30B A3B | 16.20 | 33.52 | 20.18 | 9.07 | - |
|
| 33 |
+
| Qwen3 30B A3Blimited memory | limit 8G | 10.11 | 0.18 | 6.32 | - |
|
| 34 |
+
| Gemma 3n E2B | 1G, theoretically | 36.88 | 27.06 | 12.50 | 6.66 |
|
| 35 |
+
| Gemma 3n E4B | 2G, theoretically | 21.93 | 16.58 | 7.37 | 4.01 |
|
| 36 |
+
|
| 37 |
## Model Card
|
| 38 |
|
| 39 |
<div align="center">
|