update readme

This commit is contained in:
yiranyyu
2024-05-23 18:26:23 +08:00
parent e205e7f959
commit 706063aad0

View File

@@ -25,12 +25,24 @@ In most benchmarks, MiniCPM-Llama3-V 2.5 achieves **better performance** compare
| | Phi-3-vision-128K-Instruct | MiniCPM-Llama3-V 2.5|
|:-|:----------:|:-------------------:|
| Size参数 | **4B** | 8B|
| First Token Latency首token延迟$^1$ | L: 330ms, M: 330ms, H: 330ms | **L: 48ms, M: 145ms, H: 278ms** |
| Throughtput吞吐率$^2$| 30 tokens/s | **41 tokens/s**|
| OpenCompass 2024/05 | 53.7 | **58.8** |
| OCRBench | 639.0 | **725.0**|
| RealworldQA | 58.8 | **63.5**|
| TextVQA | 72.2 | **76.6** |
| ScienceQA| **90.8** | 89.0 |
| POPE | 83.4 | **87.2** |
| MathVista | 44.5 | **54.3** |
| MMStar | 47.4 | **51.8** |
| LLaVA Bench | 64.2 | **86.7** |
| AI2D | 76.7 | **78.4** |
<small>
1: L(ow): 448pxl, M(edium): 896pxl, H(igh): 1344pxl input images.
<br>
2. Evaluation environment: A800 GPU, flash-attn=2.4.3, batch-size=1.
</small>
## Multilingual Capabilities