Upload folder using huggingface_hub
Browse files
README.md
CHANGED
|
@@ -68,15 +68,15 @@ InternVL 2.0 is a multimodal large language model series, featuring models of va
|
|
| 68 |
|
| 69 |
### Video Benchmarks
|
| 70 |
|
| 71 |
-
| Benchmark | GPT-4V |
|
| 72 |
-
| :-------------------------: | :----: |
|
| 73 |
-
| Model Size | - |
|
| 74 |
-
| | |
|
| 75 |
-
| MVBench | - |
|
| 76 |
-
| MMBench-Video<sub>8f</sub> | 1.53 |
|
| 77 |
-
| MMBench-Video<sub>16f</sub> | 1.68 |
|
| 78 |
-
| Video-MME<br>w/o subs | 59.9 |
|
| 79 |
-
| Video-MME<br>w subs | 63.3 |
|
| 80 |
|
| 81 |
- We evaluate our models on MVBench and Video-MME by extracting 16 frames from each video, and each frame was resized to a 448x448 image.
|
| 82 |
|
|
@@ -645,15 +645,15 @@ InternVL 2.0 是一个多模态大语言模型系列，包含各种规模的模
|
|
| 645 |
|
| 646 |
### 视频相关评测
|
| 647 |
|
| 648 |
-
| 评测数据集 | GPT-4V |
|
| 649 |
-
| :-------------------------: | :----: |
|
| 650 |
-
| 模型大小 | - |
|
| 651 |
-
| | |
|
| 652 |
-
| MVBench | - |
|
| 653 |
-
| MMBench-Video<sub>8f</sub> | 1.53 |
|
| 654 |
-
| MMBench-Video<sub>16f</sub> | 1.68 |
|
| 655 |
-
| Video-MME<br>w/o subs | 59.9 |
|
| 656 |
-
| Video-MME<br>w subs | 63.3 |
|
| 657 |
|
| 658 |
- 我们通过从每个视频中提取 16 帧来评估我们的模型在 MVBench 和 Video-MME 上的性能，每个视频帧被调整为 448x448 的图像。
|
| 659 |
|
|
|
|
| 68 |
|
| 69 |
### Video Benchmarks
|
| 70 |
|
| 71 |
+
| Benchmark | GPT-4V | LLaVA-NeXT-Video | InternVL-Chat-V1-5 | InternVL2-26B |
|
| 72 |
+
| :-------------------------: | :----: | :--------------: | :----------------: | :-----------: |
|
| 73 |
+
| Model Size | - | 34B | 25.5B | 25.5B |
|
| 74 |
+
| | | | | |
|
| 75 |
+
| MVBench | - | - | 52.1 | 67.5 |
|
| 76 |
+
| MMBench-Video<sub>8f</sub> | 1.53 | - | 1.26 | 1.27 |
|
| 77 |
+
| MMBench-Video<sub>16f</sub> | 1.68 | - | 1.31 | 1.41 |
|
| 78 |
+
| Video-MME<br>w/o subs | 59.9 | 52.0 | 53.6 | 54.8 |
|
| 79 |
+
| Video-MME<br>w subs | 63.3 | 54.9 | 54.5 | 57.1 |
|
| 80 |
|
| 81 |
- We evaluate our models on MVBench and Video-MME by extracting 16 frames from each video, and each frame was resized to a 448x448 image.
|
| 82 |
|
|
|
|
| 645 |
|
| 646 |
### 视频相关评测
|
| 647 |
|
| 648 |
+
| 评测数据集 | GPT-4V | LLaVA-NeXT-Video | InternVL-Chat-V1-5 | InternVL2-26B |
|
| 649 |
+
| :-------------------------: | :----: | :--------------: | :----------------: | :-----------: |
|
| 650 |
+
| 模型大小 | - | 34B | 25.5B | 25.5B |
|
| 651 |
+
| | | | | |
|
| 652 |
+
| MVBench | - | - | 52.1 | 67.5 |
|
| 653 |
+
| MMBench-Video<sub>8f</sub> | 1.53 | - | 1.26 | 1.27 |
|
| 654 |
+
| MMBench-Video<sub>16f</sub> | 1.68 | - | 1.31 | 1.41 |
|
| 655 |
+
| Video-MME<br>w/o subs | 59.9 | 52.0 | 53.6 | 54.8 |
|
| 656 |
+
| Video-MME<br>w subs | 63.3 | 54.9 | 54.5 | 57.1 |
|
| 657 |
|
| 658 |
- 我们通过从每个视频中提取 16 帧来评估我们的模型在 MVBench 和 Video-MME 上的性能，每个视频帧被调整为 448x448 的图像。
|
| 659 |
|