Upload folder using huggingface_hub
Browse files
README.md
CHANGED
|
@@ -68,13 +68,15 @@ InternVL 2.0 is a multimodal large language model series, featuring models of va
|
|
| 68 |
|
| 69 |
### Video Benchmarks
|
| 70 |
|
| 71 |
-
|
|
| 72 |
-
|
|
| 73 |
-
|
|
| 74 |
-
|
|
| 75 |
-
|
|
| 76 |
-
| Video
|
| 77 |
-
| Video
|
|
|
|
|
|
|
| 78 |
|
| 79 |
- We evaluate our models on MVBench by extracting 16 frames from each video, and each frame was resized to a 448x448 image.
|
| 80 |
|
|
@@ -453,13 +455,15 @@ InternVL 2.0 是一个多模态大语言模型系列，包含各种规模的模
|
|
| 453 |
|
| 454 |
### 视频相关评测
|
| 455 |
|
| 456 |
-
|
|
| 457 |
-
|
|
| 458 |
-
|
|
| 459 |
-
|
|
| 460 |
-
|
|
| 461 |
-
| Video
|
| 462 |
-
| Video
|
|
|
|
|
|
|
| 463 |
|
| 464 |
- 我们通过从每个视频中提取16帧来评估我们的模型在MVBench上的性能，每个视频帧被调整为448x448的图像。
|
| 465 |
|
|
|
|
| 68 |
|
| 69 |
### Video Benchmarks
|
| 70 |
|
| 71 |
+
| Benchmark | VideoChat2-Phi3 | Mini-InternVL-2B-1-5 | InternVL2-2B | InternVL2-1B |
|
| 72 |
+
| :-------------------------: | :-------------: | :------------------: | :----------: | :----------: |
|
| 73 |
+
| Model Size | 4B | 2.2B | 2.2B | 0.9B |
|
| 74 |
+
| | | | | |
|
| 75 |
+
| MVBench | 55.1 | 37.0 | 60.2 | 57.9 |
|
| 76 |
+
| MMBench-Video<sub>8f</sub> | - | 0.99 | 0.97 | 0.95 |
|
| 77 |
+
| MMBench-Video<sub>16f</sub> | - | 1.04 | 1.03 | 0.98 |
|
| 78 |
+
| Video-MME<br>wo subs | - | TODO | TODO | TODO |
|
| 79 |
+
| Video-MME<br>w/ subs | - | TODO | TODO | TODO |
|
| 80 |
|
| 81 |
- We evaluate our models on MVBench by extracting 16 frames from each video, and each frame was resized to a 448x448 image.
|
| 82 |
|
|
|
|
| 455 |
|
| 456 |
### 视频相关评测
|
| 457 |
|
| 458 |
+
| 评测数据集 | VideoChat2-Phi3 | Mini-InternVL-2B-1-5 | InternVL2-2B | InternVL2-1B |
|
| 459 |
+
| :-------------------------: | :-------------: | :------------------: | :----------: | :----------: |
|
| 460 |
+
| 模型大小 | 4B | 2.2B | 2.2B | 0.9B |
|
| 461 |
+
| | | | | |
|
| 462 |
+
| MVBench | 55.1 | 37.0 | 60.2 | 57.9 |
|
| 463 |
+
| MMBench-Video<sub>8f</sub> | - | 0.99 | 0.97 | 0.95 |
|
| 464 |
+
| MMBench-Video<sub>16f</sub> | - | 1.04 | 1.03 | 0.98 |
|
| 465 |
+
| Video-MME<br>wo subs | - | TODO | TODO | TODO |
|
| 466 |
+
| Video-MME<br>w/ subs | - | TODO | TODO | TODO |
|
| 467 |
|
| 468 |
- 我们通过从每个视频中提取16帧来评估我们的模型在MVBench上的性能，每个视频帧被调整为448x448的图像。
|
| 469 |
|