Skip to content

Commit a50399e

Browse files
xfeyxfey
committed
docs: fix homepage
1 parent a2dd7bd commit a50399e

1 file changed

Lines changed: 29 additions & 29 deletions

File tree

‎index.html‎

Lines changed: 29 additions & 29 deletions
Original file line numberDiff line numberDiff line change
@@ -292,31 +292,31 @@ <h3 class="text-xl font-bold text-slate-900" data-i18n="vl_bench_title">性能
292292
<thead>
293293
<tr>
294294
<th>Benchmark</th>
295-
<th>Qwen3-VL 4B</th>
295+
<!-- <th>Qwen3-VL 4B</th> -->
296296
<th>InternVL-3.5 4B</th>
297297
<th>UFO 8B</th>
298298
<th class="text-purple-600">Youtu-VL 4B</th>
299299
</tr>
300300
</thead>
301301
<tbody>
302302
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Visual Grounding</td></tr>
303-
<tr><td>RefCOCO val</td><td>90.7%</td><td>92.5%</td><td>91.8%</td><td class="highlight-win">93.6%</td></tr>
304-
<tr><td>RefCOCO testA</td><td>92.2%</td><td>94.3%</td><td>94.3%</td><td class="highlight-win">95.2%</td></tr>
305-
<tr><td>RefCOCO+ val</td><td>82.9%</td><td>87.6%</td><td>86.9%</td><td class="highlight-win">90.1%</td></tr>
306-
<tr><td>RefCOCOg test</td><td>87.7%</td><td>89.3%</td><td>88.6%</td><td class="highlight-win">92.9%</td></tr>
303+
<tr><td>RefCOCO val</td><td>92.5%</td><td>91.8%</td><td class="highlight-win">93.6%</td></tr>
304+
<tr><td>RefCOCO testA</td><td>94.3%</td><td>94.3%</td><td class="highlight-win">95.2%</td></tr>
305+
<tr><td>RefCOCO+ val</td><td>87.6%</td><td>86.9%</td><td class="highlight-win">90.1%</td></tr>
306+
<tr><td>RefCOCOg test</td><td>89.3%</td><td>88.6%</td><td class="highlight-win">92.9%</td></tr>
307307
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Semantic Segmentation</td></tr>
308-
<tr><td>ADE20k</td><td>×</td><td>×</td><td>54.5%</td><td class="highlight-second">54.2%</td></tr>
309-
<tr><td>Cityscapes</td><td>×</td><td>×</td><td>-</td><td class="highlight-win">70.4%</td></tr>
310-
<tr><td>VOC20</td><td>×</td><td>×</td><td>-</td><td class="highlight-win">92.5%</td></tr>
308+
<tr><td>ADE20k</td><td>×</td><td>54.5%</td><td class="highlight-second">54.2%</td></tr>
309+
<tr><td>Cityscapes</td><td>×</td><td>-</td><td class="highlight-win">70.4%</td></tr>
310+
<tr><td>VOC20</td><td>×</td><td>-</td><td class="highlight-win">92.5%</td></tr>
311311
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Referring Segmentation</td></tr>
312-
<tr><td>RefCOCO val</td><td>×</td><td>×</td><td>80.0%</td><td class="highlight-win">80.7%</td></tr>
313-
<tr><td>RefCOCO testA</td><td>×</td><td>×</td><td>81.6%</td><td class="highlight-win">82.0%</td></tr>
312+
<tr><td>RefCOCO val</td><td>×</td><td>80.0%</td><td class="highlight-win">80.7%</td></tr>
313+
<tr><td>RefCOCO testA</td><td>×</td><td>81.6%</td><td class="highlight-win">82.0%</td></tr>
314314
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Depth Estimation</td></tr>
315-
<tr><td>NYUv2 (δ1)</td><td>×</td><td>×</td><td>93.6%</td><td class="highlight-second">90.4%</td></tr>
316-
<tr><td>Cityscapes</td><td>×</td><td>×</td><td>-</td><td class="highlight-win">92.7%</td></tr>
315+
<tr><td>NYUv2 (δ1)</td><td>×</td><td>93.6%</td><td class="highlight-second">90.4%</td></tr>
316+
<tr><td>Cityscapes</td><td>×</td><td>-</td><td class="highlight-win">92.7%</td></tr>
317317
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Object Counting</td></tr>
318-
<tr><td>TallyQA-Simple</td><td>79.0%</td><td>77.6%</td><td>×</td><td class="highlight-win">85.1%</td></tr>
319-
<tr><td>CountBench</td><td>78.4%</td><td>79.4%</td><td>×</td><td class="highlight-win">88.6%</td></tr>
318+
<tr><td>TallyQA-Simple</td><td>77.6%</td><td>×</td><td class="highlight-win">85.1%</td></tr>
319+
<tr><td>CountBench</td><td>79.4%</td><td>×</td><td class="highlight-win">88.6%</td></tr>
320320
</tbody>
321321
</table>
322322
<p class="text-xs text-slate-500 mt-4 text-center" data-i18n="vl_table_note1">* × 表示模型不支持该任务。</p>
@@ -330,30 +330,30 @@ <h3 class="text-xl font-bold text-slate-900" data-i18n="vl_bench_title">性能
330330
<th>Benchmark</th>
331331
<th>Qwen3-VL 8B</th>
332332
<th>InternVL-3.5 4B</th>
333-
<th>Qwen3-VL 4B</th>
333+
<!-- <th>Qwen3-VL 4B</th> -->
334334
<th class="text-purple-600">Youtu-VL 4B</th>
335335
</tr>
336336
</thead>
337337
<tbody>
338338
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">General VQA</td></tr>
339-
<tr><td>MMBench_EN</td><td>84.5%</td><td>80.3%</td><td>83.9%</td><td class="highlight-second">83.9%</td></tr>
340-
<tr><td>MMStar</td><td>70.9%</td><td>65.0%</td><td>69.8%</td><td class="highlight-win">71.1%</td></tr>
341-
<tr><td>MME (/2800)</td><td>-</td><td>2272</td><td>2309</td><td class="highlight-win">2384</td></tr>
342-
<tr><td>ScienceQA_val</td><td>-</td><td>-</td><td>94.7%</td><td class="highlight-win">97.0%</td></tr>
339+
<tr><td>MMBench_EN</td><td>84.5%</td><td>80.3%</td><td class="highlight-second">83.9%</td></tr>
340+
<tr><td>MMStar</td><td>70.9%</td><td>65.0%</td><td class="highlight-win">71.1%</td></tr>
341+
<tr><td>MME (/2800)</td><td>-</td><td>2272</td><td class="highlight-win">2384</td></tr>
342+
<tr><td>ScienceQA_val</td><td>-</td><td>-</td><td class="highlight-win">97.0%</td></tr>
343343
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Multimodal Reasoning</td></tr>
344-
<tr><td>VisuLogic</td><td>22.5%</td><td>-</td><td>19.0%</td><td class="highlight-win">25.7%</td></tr>
345-
<tr><td>MathVista_mini</td><td>77.2%</td><td>77.1%</td><td>73.7%</td><td class="highlight-second">76.5%</td></tr>
346-
<tr><td>MathVerse_mini</td><td>62.1%</td><td>45.8%</td><td>46.8%</td><td class="highlight-second">56.5%</td></tr>
347-
<tr><td>VLMsAreBlind</td><td>74.0%</td><td>-</td><td>71.9%</td><td class="highlight-win">88.9%</td></tr>
344+
<tr><td>VisuLogic</td><td>22.5%</td><td>-</td>><td class="highlight-win">25.7%</td></tr>
345+
<tr><td>MathVista_mini</td><td>77.2%</td><td>77.1%</td><td class="highlight-second">76.5%</td></tr>
346+
<tr><td>MathVerse_mini</td><td>62.1%</td><td>45.8%</td><td class="highlight-second">56.5%</td></tr>
347+
<tr><td>VLMsAreBlind</td><td>74.0%</td><td>-</td><td class="highlight-win">88.9%</td></tr>
348348
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">OCR & Document</td></tr>
349-
<tr><td>AI2D_test</td><td>85.7%</td><td>82.6%</td><td>84.1%</td><td class="highlight-second">85.6%</td></tr>
350-
<tr><td>DocVQA_val</td><td>96.1%</td><td>92.4%</td><td>95.3%</td><td class="highlight-second">94.4%</td></tr>
351-
<tr><td>ChartQA_test</td><td>89.6%</td><td>86.0%</td><td>84.6%</td><td class="highlight-second">85.3%</td></tr>
349+
<tr><td>AI2D_test</td><td>85.7%</td><td>82.6%</td><td class="highlight-second">85.6%</td></tr>
350+
<tr><td>DocVQA_val</td><td>96.1%</td><td>92.4%</td><td class="highlight-second">94.4%</td></tr>
351+
<tr><td>ChartQA_test</td><td>89.6%</td><td>86.0%</td><td class="highlight-second">85.3%</td></tr>
352352
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">GUI Agent</td></tr>
353-
<tr><td>ScreenSpot Pro</td><td>54.6%</td><td>-</td><td>59.5%</td><td class="highlight-win">59.6%</td></tr>
354-
<tr><td>OSWorld</td><td>33.9%</td><td>-</td><td>26.2%</td><td class="highlight-win">38.8%</td></tr>
353+
<tr><td>ScreenSpot Pro</td><td>54.6%</td><td>-</td><td class="highlight-win">59.6%</td></tr>
354+
<tr><td>OSWorld</td><td>33.9%</td><td>-</td><td class="highlight-win">38.8%</td></tr>
355355
<tr><td colspan="5" class="font-semibold bg-slate-50 text-slate-700">Real-World</td></tr>
356-
<tr><td>RealWorldQA</td><td>71.5%</td><td>66.3%</td><td>70.9%</td><td class="highlight-win">74.6%</td></tr>
356+
<tr><td>RealWorldQA</td><td>71.5%</td><td>66.3%</td><td class="highlight-win">74.6%</td></tr>
357357
</tbody>
358358
</table>
359359
<p class="text-xs text-slate-500 mt-4 text-center" data-i18n="vl_table_note2">* 对比同级别 VLM 指令模型表现。</p>

0 commit comments

Comments
 (0)