Update README.md
Browse files
README.md
CHANGED
|
@@ -87,10 +87,21 @@ model-index:
|
|
| 87 |
|
| 88 |
|
| 89 |
## 📈 Performance
|
|
|
|
|
|
|
| 90 |
| Model | MVBench | LongVideoBench | VideoMME(w/o sub)|
|
| 91 |
| --- | --- | --- | --- |
|
| 92 |
|InternVideo2.5| 75.7 | 60.6 | 65.1|
|
| 93 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
## 🚀 How to use the model
|
| 95 |
|
| 96 |
First, you need to install [flash attention2](https://github.com/Dao-AILab/flash-attention) and some other modules. We provide a simple installation example below:
|
|
|
|
| 87 |
|
| 88 |
|
| 89 |
## 📈 Performance
|
| 90 |
+
|
| 91 |
+
- VideoBenchmark
|
| 92 |
| Model | MVBench | LongVideoBench | VideoMME(w/o sub)|
|
| 93 |
| --- | --- | --- | --- |
|
| 94 |
|InternVideo2.5| 75.7 | 60.6 | 65.1|
|
| 95 |
|
| 96 |
+
- Inference Speed
|
| 97 |
+
|
| 98 |
+
We measured the average inference speed (tokens/s) of generating 1024 new tokens and 5198 (8192-2998) tokens with the context of an video (which takes 2998 tokens) under BF16 precision.
|
| 99 |
+
|
| 100 |
+
|Quantization | Speed (3022 tokens) | Speed (8192 tokens)|
|
| 101 |
+
|--- |--- |---|
|
| 102 |
+
|BF16 | 33.40 | 31.91 |
|
| 103 |
+
|
| 104 |
+
|
| 105 |
## 🚀 How to use the model
|
| 106 |
|
| 107 |
First, you need to install [flash attention2](https://github.com/Dao-AILab/flash-attention) and some other modules. We provide a simple installation example below:
|