Spaces:
Running
Running
temp fix for the leaderboard
Browse files- app.py +14 -5
- archive/page_old_scores.jsonl +0 -3
- datasets.py +1 -1
- rankings/mmeb_ranking.csv +41 -41
- rankings/mmeb_ranking.jsonl +41 -41
- rankings/visdoc_ranking.csv +43 -43
- rankings/visdoc_ranking.jsonl +42 -42
- utils_v2.py +15 -10
app.py
CHANGED
|
@@ -149,23 +149,32 @@ with gr.Blocks() as block:
|
|
| 149 |
# table 4, visual document scores only
|
| 150 |
with gr.TabItem("📑 Visual Doc", elem_id="tab-visdoc", id=4):
|
| 151 |
gr.Markdown(v2.TABLE_INTRODUCTION_D)
|
| 152 |
-
df2_d = v2.rank_models(df2[v2.COLUMN_NAMES_D], 'Visdoc-Overall')
|
| 153 |
# ================= HARD CODED TEMPORARY FIX =================
|
| 154 |
temp_df_vd = pd.read_json('archive/page_old_scores.jsonl', orient='records', lines=True)
|
| 155 |
df2_d = df2_d.merge(temp_df_vd, on='Models', how='left')
|
| 156 |
def special_process_visdoc(row):
|
| 157 |
if not pd.isna(row['ViDoSeek-page-old']):
|
| 158 |
-
row['ViDoSeek-page'] = '
|
| 159 |
if not pd.isna(row['MMLongBench-page-old']):
|
| 160 |
-
row['MMLongBench-page'] = '
|
| 161 |
return row
|
| 162 |
df2_d = df2_d.apply(special_process_visdoc, axis=1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 163 |
# ==========================================================
|
| 164 |
data_component5 = gr.components.Dataframe(
|
| 165 |
value=df2_d,
|
| 166 |
-
headers=
|
| 167 |
type="pandas",
|
| 168 |
-
datatype=
|
| 169 |
interactive=False,
|
| 170 |
visible=True,
|
| 171 |
max_height=2400,
|
|
|
|
| 149 |
# table 4, visual document scores only
|
| 150 |
with gr.TabItem("📑 Visual Doc", elem_id="tab-visdoc", id=4):
|
| 151 |
gr.Markdown(v2.TABLE_INTRODUCTION_D)
|
| 152 |
+
df2_d = v2.rank_models(df2[v2.COLUMN_NAMES_D+['ViDoSeek-page', 'MMLongBench-page']], 'Visdoc-Overall')
|
| 153 |
# ================= HARD CODED TEMPORARY FIX =================
|
| 154 |
temp_df_vd = pd.read_json('archive/page_old_scores.jsonl', orient='records', lines=True)
|
| 155 |
df2_d = df2_d.merge(temp_df_vd, on='Models', how='left')
|
| 156 |
def special_process_visdoc(row):
|
| 157 |
if not pd.isna(row['ViDoSeek-page-old']):
|
| 158 |
+
row['ViDoSeek-page'] = '⚠️! Please fix this score!'
|
| 159 |
if not pd.isna(row['MMLongBench-page-old']):
|
| 160 |
+
row['MMLongBench-page'] = '⚠️! Please fix this score!'
|
| 161 |
return row
|
| 162 |
df2_d = df2_d.apply(special_process_visdoc, axis=1)
|
| 163 |
+
df2_d = df2_d.rename(columns={
|
| 164 |
+
'ViDoSeek-page': 'ViDoSeek-page-fixed',
|
| 165 |
+
'MMLongBench-page': 'MMLongBench-page-fixed'
|
| 166 |
+
})
|
| 167 |
+
temp_col_names_d = v2.COLUMN_NAMES_D + ['ViDoSeek-page-fixed', 'MMLongBench-page-fixed', 'ViDoSeek-page-old', 'MMLongBench-page-old']
|
| 168 |
+
print(temp_col_names_d)
|
| 169 |
+
temp_data_type_d = v2.DATA_TITLE_TYPE_D + ['number', 'number']
|
| 170 |
+
df2_d[['ViDoSeek-page-old', 'MMLongBench-page-old']] = df2_d[['ViDoSeek-page-old', 'MMLongBench-page-old']].fillna('✅')
|
| 171 |
+
df2_d = df2_d[temp_col_names_d]
|
| 172 |
# ==========================================================
|
| 173 |
data_component5 = gr.components.Dataframe(
|
| 174 |
value=df2_d,
|
| 175 |
+
headers=temp_col_names_d,
|
| 176 |
type="pandas",
|
| 177 |
+
datatype=temp_data_type_d,
|
| 178 |
interactive=False,
|
| 179 |
visible=True,
|
| 180 |
max_height=2400,
|
archive/page_old_scores.jsonl
CHANGED
|
@@ -1,7 +1,4 @@
|
|
| 1 |
-
{"Models":"Taichu-UniRetriever-8B","ViDoSeek-page-old":92.42,"MMLongBench-page-old":63.78}
|
| 2 |
-
{"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-8B\">Qwen3-VL-Embedding-8B<\/a>","ViDoSeek-page-old":88.38,"MMLongBench-page-old":60.29}
|
| 3 |
{"Models":"<a href=\"https:\/\/interestfm-tte.github.io\/\">IFM-TTE-7B<\/a>","ViDoSeek-page-old":50.29,"MMLongBench-page-old":28.18}
|
| 4 |
-
{"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-2B\">Qwen3-VL-Embedding-2B<\/a>","ViDoSeek-page-old":81.94,"MMLongBench-page-old":55.73}
|
| 5 |
{"Models":"<a href=\"https:\/\/console.volcengine.com\/ark\/region:ark+cn-beijing\/model\/detail?Id=doubao-embedding-vision\">seed1.6-embedding-1215<\/a>","ViDoSeek-page-old":22.29,"MMLongBench-page-old":17.15}
|
| 6 |
{"Models":"WeMM-Embedding-8B","ViDoSeek-page-old":23.34,"MMLongBench-page-old":17.05}
|
| 7 |
{"Models":"<a href=\"https:\/\/github.com\/360CVGroup\/RzenEmbed\">RzenEmbed-v2-7B<\/a>","ViDoSeek-page-old":23.09,"MMLongBench-page-old":15.62}
|
|
|
|
|
|
|
|
|
|
| 1 |
{"Models":"<a href=\"https:\/\/interestfm-tte.github.io\/\">IFM-TTE-7B<\/a>","ViDoSeek-page-old":50.29,"MMLongBench-page-old":28.18}
|
|
|
|
| 2 |
{"Models":"<a href=\"https:\/\/console.volcengine.com\/ark\/region:ark+cn-beijing\/model\/detail?Id=doubao-embedding-vision\">seed1.6-embedding-1215<\/a>","ViDoSeek-page-old":22.29,"MMLongBench-page-old":17.15}
|
| 3 |
{"Models":"WeMM-Embedding-8B","ViDoSeek-page-old":23.34,"MMLongBench-page-old":17.05}
|
| 4 |
{"Models":"<a href=\"https:\/\/github.com\/360CVGroup\/RzenEmbed\">RzenEmbed-v2-7B<\/a>","ViDoSeek-page-old":23.09,"MMLongBench-page-old":15.62}
|
datasets.py
CHANGED
|
@@ -18,7 +18,7 @@ DATASETS = {
|
|
| 18 |
"ViDoRe-V1": ['ViDoRe_arxivqa', 'ViDoRe_docvqa', 'ViDoRe_infovqa', 'ViDoRe_tabfquad', 'ViDoRe_tatdqa', 'ViDoRe_shiftproject', 'ViDoRe_syntheticDocQA_artificial_intelligence', 'ViDoRe_syntheticDocQA_energy', 'ViDoRe_syntheticDocQA_government_reports', 'ViDoRe_syntheticDocQA_healthcare_industry'],
|
| 19 |
"ViDoRe-V2": ["ViDoRe_esg_reports_human_labeled_v2","ViDoRe_biomedical_lectures_v2_multilingual", "ViDoRe_economics_reports_v2_multilingual", "ViDoRe_esg_reports_v2_multilingual"], # "ViDoRe_biomedical_lectures_v2", "ViDoRe_economics_reports_v2", "ViDoRe_esg_reports_v2"
|
| 20 |
"VisRAG": ['VisRAG_ArxivQA', 'VisRAG_ChartQA', 'VisRAG_MP-DocVQA', 'VisRAG_SlideVQA', 'VisRAG_InfoVQA', 'VisRAG_PlotQA'],
|
| 21 |
-
"VisDoc-OOD": ['ViDoSeek-
|
| 22 |
},
|
| 23 |
"video": {
|
| 24 |
"V-CLS": ['K700', 'UCF101', 'HMDB51', 'SmthSmthV2', 'Breakfast'],
|
|
|
|
| 18 |
"ViDoRe-V1": ['ViDoRe_arxivqa', 'ViDoRe_docvqa', 'ViDoRe_infovqa', 'ViDoRe_tabfquad', 'ViDoRe_tatdqa', 'ViDoRe_shiftproject', 'ViDoRe_syntheticDocQA_artificial_intelligence', 'ViDoRe_syntheticDocQA_energy', 'ViDoRe_syntheticDocQA_government_reports', 'ViDoRe_syntheticDocQA_healthcare_industry'],
|
| 19 |
"ViDoRe-V2": ["ViDoRe_esg_reports_human_labeled_v2","ViDoRe_biomedical_lectures_v2_multilingual", "ViDoRe_economics_reports_v2_multilingual", "ViDoRe_esg_reports_v2_multilingual"], # "ViDoRe_biomedical_lectures_v2", "ViDoRe_economics_reports_v2", "ViDoRe_esg_reports_v2"
|
| 20 |
"VisRAG": ['VisRAG_ArxivQA', 'VisRAG_ChartQA', 'VisRAG_MP-DocVQA', 'VisRAG_SlideVQA', 'VisRAG_InfoVQA', 'VisRAG_PlotQA'],
|
| 21 |
+
"VisDoc-OOD": ['ViDoSeek-doc', 'MMLongBench-doc'] # 'ViDoSeek-page', 'MMLongBench-page'
|
| 22 |
},
|
| 23 |
"video": {
|
| 24 |
"V-CLS": ['K700', 'UCF101', 'HMDB51', 'SmthSmthV2', 'Breakfast'],
|
rankings/mmeb_ranking.csv
CHANGED
|
@@ -1,43 +1,43 @@
|
|
| 1 |
Rank,Models,Model Size(B),Date,Overall,Image-Overall,Video-Overall,Visdoc-Overall
|
| 2 |
-
1,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-8B"">Qwen3-VL-Embedding-8B</a>",8.14,2026-01-06,
|
| 3 |
-
2,"<a href=""https://console.volcengine.com/ark/region:ark+cn-beijing/model/detail?Id=doubao-embedding-vision"">seed1.6-embedding-1215</a>",unknown,2025-12-18,
|
| 4 |
-
3,WeMM-Embedding-8B,8.77,2025-12-16,
|
| 5 |
-
4,"<a href=""https://interestfm-tte.github.io/"">IFM-TTE-7B</a>",8.29,2025-10-11,
|
| 6 |
-
5,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-2B"">Qwen3-VL-Embedding-2B</a>",2.13,2026-01-06,
|
| 7 |
-
6,WeMM-Embedding-2B,2.13,2025-12-16,
|
| 8 |
-
7,"<a href=""https://github.com/360CVGroup/RzenEmbed"">RzenEmbed-v2-7B</a>",8.29,2025-09-20,
|
| 9 |
-
8,"<a href=""https://seed1-6-embedding.github.io"">seed-1.6-embedding</a>",unknown,2025-06-17,
|
| 10 |
-
9,RzenEmbed-v1-7B,8.29,2025-07-30,
|
| 11 |
-
10,"<a href=""https://huggingface.co/OpenSearch-AI/Ops-MM-embedding-v1-7B"">Ops-MM-embedding-v1-7B</a>",8.29,2025-07-02,
|
| 12 |
-
11,"<a href=""https://huggingface.co/Haon-Chen/e5-omni-7B"">e5-omni-7B</a>",8.0,2026-01-09,
|
| 13 |
-
12,Crotchet-embedding-2B,2.13,2025-12-31,
|
| 14 |
-
13,"<a href=""https://huggingface.co/zhibinlan/UME-R1-7B"">UME-R1-7B</a>",8.29,2025-10-20,
|
| 15 |
-
14,RzenEmbed-v1-2B,2.21,2025-07-16,
|
| 16 |
-
15,"<a href=""https://huggingface.co/OpenSearch-AI/Ops-MM-embedding-v1-2B"">Ops-MM-embedding-v1-2B</a>",2.21,2025-07-02,
|
| 17 |
-
16,"<a href=""https://huggingface.co/Haon-Chen/e5-omni-3B"">e5-omni-3B</a>",3.0,2026-01-09,
|
| 18 |
-
17,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-7B</a>",8.03,2025-07-06,
|
| 19 |
-
18,"<a href=""https://huggingface.co/zhibinlan/UME-R1-2B"">UME-R1-2B</a>",2.21,2025-10-20,
|
| 20 |
-
19,"<a href=""https://github.com/GaryGuTC/UniME-v2"">UniME-V2-LLaVA-OneVision-7B</a>",8.03,2025-10-15,
|
| 21 |
-
20,"<a href=""https://huggingface.co/VLM2Vec/VLM2Vec-V2.0"">VLM2Vec-V2.0-Qwen2VL-2B</a>",2.21,2025-12-29,
|
| 22 |
-
21,"<a href=""https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-7B-Instruct"">gme-Qwen2-VL-7B-Instruct</a>",8.29,2025-12-29,
|
| 23 |
-
22,"<a href=""https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-2B-Instruct"">gme-Qwen2-VL-2B-Instruct</a>",2.21,2025-12-29,
|
| 24 |
-
23,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-0.5B</a>",0.894,2025-07-06,
|
| 25 |
-
24,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-7B"">VLM2Vec-V1-Qwen2VL-7B</a>",8.29,2025-12-29,
|
| 26 |
-
25,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret-Qwen2.5VL-7b"">LamRA-Ret-Qwen2.5VL-7b</a>",8.29,2025-12-29,
|
| 27 |
-
26,colpali-v1.3,unknown,2025-12-29,
|
| 28 |
-
27,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-2B"">VLM2Vec-V1-Qwen2VL-2B</a>",2.21,2025-12-29,
|
| 29 |
-
28,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret"">LamRA-Ret</a>",8.29,2025-12-29,
|
| 30 |
-
29,QQMM-embed-v3,8.29,2025-12-30,
|
| 31 |
-
30,"<a href=""https://github.com/QQ-MM/QQMM-embed"">QQMM-embed-v2</a>",8.29,2025-09-15,
|
| 32 |
-
31,UniVec-CoT-7B,8.29,unknown,
|
| 33 |
-
32,OEmbedding-v1-7B,8.29,2025-10-14,
|
| 34 |
-
33,ReCo-7B,8.29,2025-08-15,34.
|
| 35 |
-
34,UniVec-7B,8.29,unknown,34.
|
| 36 |
-
35,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-7B</a>",8.29,2025-11-24,34.
|
| 37 |
-
36,dp-embedding-v3-lite,8.29,2025-12-15,
|
| 38 |
-
37,TCE-v1,8.0,2025-10-31,
|
| 39 |
-
38,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-3B</a>",3.75,2025-11-24,
|
| 40 |
-
39,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-2B</a>",2.21,2025-11-24,
|
| 41 |
-
40,Taichu-UniRetriever-8B,8.77,2026-01-15,
|
| 42 |
-
41,Taichu-UniRetriever-v1-2B,2.13,2026-01-06,
|
| 43 |
42,TCR,8.77,2026-01-15,0.0,0.0,0.0,0.0
|
|
|
|
| 1 |
Rank,Models,Model Size(B),Date,Overall,Image-Overall,Video-Overall,Visdoc-Overall
|
| 2 |
+
1,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-8B"">Qwen3-VL-Embedding-8B</a>",8.14,2026-01-06,79.87,80.12,67.15,83.09
|
| 3 |
+
2,"<a href=""https://console.volcengine.com/ark/region:ark+cn-beijing/model/detail?Id=doubao-embedding-vision"">seed1.6-embedding-1215</a>",unknown,2025-12-18,77.58,77.99,67.74,83.16
|
| 4 |
+
3,WeMM-Embedding-8B,8.77,2025-12-16,76.38,78.09,63.24,82.49
|
| 5 |
+
4,"<a href=""https://interestfm-tte.github.io/"">IFM-TTE-7B</a>",8.29,2025-10-11,76.02,77.9,59.19,83.14
|
| 6 |
+
5,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-2B"">Qwen3-VL-Embedding-2B</a>",2.13,2026-01-06,75.18,74.96,61.87,80.16
|
| 7 |
+
6,WeMM-Embedding-2B,2.13,2025-12-16,73.56,76.08,58.67,79.85
|
| 8 |
+
7,"<a href=""https://github.com/360CVGroup/RzenEmbed"">RzenEmbed-v2-7B</a>",8.29,2025-09-20,73.5,75.92,55.73,82.31
|
| 9 |
+
8,"<a href=""https://seed1-6-embedding.github.io"">seed-1.6-embedding</a>",unknown,2025-06-17,73.14,77.78,55.34,78.37
|
| 10 |
+
9,RzenEmbed-v1-7B,8.29,2025-07-30,70.69,73.6,48.87,82.0
|
| 11 |
+
10,"<a href=""https://huggingface.co/OpenSearch-AI/Ops-MM-embedding-v1-7B"">Ops-MM-embedding-v1-7B</a>",8.29,2025-07-02,69.39,72.72,53.76,74.99
|
| 12 |
+
11,"<a href=""https://huggingface.co/Haon-Chen/e5-omni-7B"">e5-omni-7B</a>",8.0,2026-01-09,68.1,71.23,43.54,81.3
|
| 13 |
+
12,Crotchet-embedding-2B,2.13,2025-12-31,67.3,68.18,50.63,73.59
|
| 14 |
+
13,"<a href=""https://huggingface.co/zhibinlan/UME-R1-7B"">UME-R1-7B</a>",8.29,2025-10-20,66.2,71.25,47.5,71.71
|
| 15 |
+
14,RzenEmbed-v1-2B,2.21,2025-07-16,66.05,68.53,42.62,79.4
|
| 16 |
+
15,"<a href=""https://huggingface.co/OpenSearch-AI/Ops-MM-embedding-v1-2B"">Ops-MM-embedding-v1-2B</a>",2.21,2025-07-02,65.11,69.03,47.56,71.48
|
| 17 |
+
16,"<a href=""https://huggingface.co/Haon-Chen/e5-omni-3B"">e5-omni-3B</a>",3.0,2026-01-09,64.74,67.59,40.63,78.11
|
| 18 |
+
17,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-7B</a>",8.03,2025-07-06,62.23,67.56,42.4,68.1
|
| 19 |
+
18,"<a href=""https://huggingface.co/zhibinlan/UME-R1-2B"">UME-R1-2B</a>",2.21,2025-10-20,61.69,66.56,42.23,68.16
|
| 20 |
+
19,"<a href=""https://github.com/GaryGuTC/UniME-v2"">UniME-V2-LLaVA-OneVision-7B</a>",8.03,2025-10-15,61.13,71.77,39.01,60.24
|
| 21 |
+
20,"<a href=""https://huggingface.co/VLM2Vec/VLM2Vec-V2.0"">VLM2Vec-V2.0-Qwen2VL-2B</a>",2.21,2025-12-29,59.01,64.92,34.67,63.56
|
| 22 |
+
21,"<a href=""https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-7B-Instruct"">gme-Qwen2-VL-7B-Instruct</a>",8.29,2025-12-29,58.83,55.93,38.43,73.96
|
| 23 |
+
22,"<a href=""https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-2B-Instruct"">gme-Qwen2-VL-2B-Instruct</a>",2.21,2025-12-29,55.06,51.87,33.59,71.88
|
| 24 |
+
23,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-0.5B</a>",0.894,2025-07-06,50.99,55.43,35.87,54.83
|
| 25 |
+
24,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-7B"">VLM2Vec-V1-Qwen2VL-7B</a>",8.29,2025-12-29,48.23,65.37,33.79,26.82
|
| 26 |
+
25,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret-Qwen2.5VL-7b"">LamRA-Ret-Qwen2.5VL-7b</a>",8.29,2025-12-29,48.05,52.54,33.83,47.61
|
| 27 |
+
26,colpali-v1.3,unknown,2025-12-29,45.36,34.89,28.23,70.5
|
| 28 |
+
27,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-2B"">VLM2Vec-V1-Qwen2VL-2B</a>",2.21,2025-12-29,43.81,59.73,28.55,26.0
|
| 29 |
+
28,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret"">LamRA-Ret</a>",8.29,2025-12-29,42.7,54.13,35.21,27.72
|
| 30 |
+
29,QQMM-embed-v3,8.29,2025-12-30,37.05,78.21,0.0,0.0
|
| 31 |
+
30,"<a href=""https://github.com/QQ-MM/QQMM-embed"">QQMM-embed-v2</a>",8.29,2025-09-15,35.66,75.28,0.0,0.0
|
| 32 |
+
31,UniVec-CoT-7B,8.29,unknown,35.1,74.11,0.0,0.0
|
| 33 |
+
32,OEmbedding-v1-7B,8.29,2025-10-14,35.08,74.05,0.0,0.0
|
| 34 |
+
33,ReCo-7B,8.29,2025-08-15,34.99,73.87,0.0,0.0
|
| 35 |
+
34,UniVec-7B,8.29,unknown,34.96,73.79,0.0,0.0
|
| 36 |
+
35,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-7B</a>",8.29,2025-11-24,34.91,73.7,0.0,0.0
|
| 37 |
+
36,dp-embedding-v3-lite,8.29,2025-12-15,34.76,73.38,0.0,0.0
|
| 38 |
+
37,TCE-v1,8.0,2025-10-31,34.27,72.36,0.0,0.0
|
| 39 |
+
38,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-3B</a>",3.75,2025-11-24,33.25,70.2,0.0,0.0
|
| 40 |
+
39,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-2B</a>",2.21,2025-11-24,32.78,69.21,0.0,0.0
|
| 41 |
+
40,Taichu-UniRetriever-8B,8.77,2026-01-15,26.29,0.0,0.0,83.72
|
| 42 |
+
41,Taichu-UniRetriever-v1-2B,2.13,2026-01-06,24.21,0.0,0.0,81.93
|
| 43 |
42,TCR,8.77,2026-01-15,0.0,0.0,0.0,0.0
|
rankings/mmeb_ranking.jsonl
CHANGED
|
@@ -1,42 +1,42 @@
|
|
| 1 |
-
{"Rank":1,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-8B\">Qwen3-VL-Embedding-8B<\/a>","Model Size(B)":8.14,"Date":"2026-01-06","Overall":
|
| 2 |
-
{"Rank":2,"Models":"<a href=\"https:\/\/console.volcengine.com\/ark\/region:ark+cn-beijing\/model\/detail?Id=doubao-embedding-vision\">seed1.6-embedding-1215<\/a>","Model Size(B)":"unknown","Date":"2025-12-18","Overall":
|
| 3 |
-
{"Rank":3,"Models":"WeMM-Embedding-8B","Model Size(B)":8.77,"Date":"2025-12-16","Overall":
|
| 4 |
-
{"Rank":4,"Models":"<a href=\"https:\/\/interestfm-tte.github.io\/\">IFM-TTE-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-11","Overall":
|
| 5 |
-
{"Rank":5,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-2B\">Qwen3-VL-Embedding-2B<\/a>","Model Size(B)":2.13,"Date":"2026-01-06","Overall":
|
| 6 |
-
{"Rank":6,"Models":"WeMM-Embedding-2B","Model Size(B)":2.13,"Date":"2025-12-16","Overall":
|
| 7 |
-
{"Rank":7,"Models":"<a href=\"https:\/\/github.com\/360CVGroup\/RzenEmbed\">RzenEmbed-v2-7B<\/a>","Model Size(B)":8.29,"Date":"2025-09-20","Overall":
|
| 8 |
-
{"Rank":8,"Models":"<a href=\"https:\/\/seed1-6-embedding.github.io\">seed-1.6-embedding<\/a>","Model Size(B)":"unknown","Date":"2025-06-17","Overall":
|
| 9 |
-
{"Rank":9,"Models":"RzenEmbed-v1-7B","Model Size(B)":8.29,"Date":"2025-07-30","Overall":
|
| 10 |
-
{"Rank":10,"Models":"<a href=\"https:\/\/huggingface.co\/OpenSearch-AI\/Ops-MM-embedding-v1-7B\">Ops-MM-embedding-v1-7B<\/a>","Model Size(B)":8.29,"Date":"2025-07-02","Overall":
|
| 11 |
-
{"Rank":11,"Models":"<a href=\"https:\/\/huggingface.co\/Haon-Chen\/e5-omni-7B\">e5-omni-7B<\/a>","Model Size(B)":8.0,"Date":"2026-01-09","Overall":
|
| 12 |
-
{"Rank":12,"Models":"Crotchet-embedding-2B","Model Size(B)":2.13,"Date":"2025-12-31","Overall":
|
| 13 |
-
{"Rank":13,"Models":"<a href=\"https:\/\/huggingface.co\/zhibinlan\/UME-R1-7B\">UME-R1-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-20","Overall":
|
| 14 |
-
{"Rank":14,"Models":"RzenEmbed-v1-2B","Model Size(B)":2.21,"Date":"2025-07-16","Overall":
|
| 15 |
-
{"Rank":15,"Models":"<a href=\"https:\/\/huggingface.co\/OpenSearch-AI\/Ops-MM-embedding-v1-2B\">Ops-MM-embedding-v1-2B<\/a>","Model Size(B)":2.21,"Date":"2025-07-02","Overall":
|
| 16 |
-
{"Rank":16,"Models":"<a href=\"https:\/\/huggingface.co\/Haon-Chen\/e5-omni-3B\">e5-omni-3B<\/a>","Model Size(B)":3.0,"Date":"2026-01-09","Overall":
|
| 17 |
-
{"Rank":17,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-7B<\/a>","Model Size(B)":8.03,"Date":"2025-07-06","Overall":
|
| 18 |
-
{"Rank":18,"Models":"<a href=\"https:\/\/huggingface.co\/zhibinlan\/UME-R1-2B\">UME-R1-2B<\/a>","Model Size(B)":2.21,"Date":"2025-10-20","Overall":
|
| 19 |
-
{"Rank":19,"Models":"<a href=\"https:\/\/github.com\/GaryGuTC\/UniME-v2\">UniME-V2-LLaVA-OneVision-7B<\/a>","Model Size(B)":8.03,"Date":"2025-10-15","Overall":
|
| 20 |
-
{"Rank":20,"Models":"<a href=\"https:\/\/huggingface.co\/VLM2Vec\/VLM2Vec-V2.0\">VLM2Vec-V2.0-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Overall":
|
| 21 |
-
{"Rank":21,"Models":"<a href=\"https:\/\/huggingface.co\/Alibaba-NLP\/gme-Qwen2-VL-7B-Instruct\">gme-Qwen2-VL-7B-Instruct<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":
|
| 22 |
-
{"Rank":22,"Models":"<a href=\"https:\/\/huggingface.co\/Alibaba-NLP\/gme-Qwen2-VL-2B-Instruct\">gme-Qwen2-VL-2B-Instruct<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Overall":
|
| 23 |
-
{"Rank":23,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-0.5B<\/a>","Model Size(B)":0.894,"Date":"2025-07-06","Overall":
|
| 24 |
-
{"Rank":24,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-7B\">VLM2Vec-V1-Qwen2VL-7B<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":
|
| 25 |
-
{"Rank":25,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret-Qwen2.5VL-7b\">LamRA-Ret-Qwen2.5VL-7b<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":
|
| 26 |
-
{"Rank":26,"Models":"colpali-v1.3","Model Size(B)":"unknown","Date":"2025-12-29","Overall":
|
| 27 |
-
{"Rank":27,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-2B\">VLM2Vec-V1-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Overall":
|
| 28 |
-
{"Rank":28,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret\">LamRA-Ret<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":
|
| 29 |
-
{"Rank":29,"Models":"QQMM-embed-v3","Model Size(B)":8.29,"Date":"2025-12-30","Overall":
|
| 30 |
-
{"Rank":30,"Models":"<a href=\"https:\/\/github.com\/QQ-MM\/QQMM-embed\">QQMM-embed-v2<\/a>","Model Size(B)":8.29,"Date":"2025-09-15","Overall":
|
| 31 |
-
{"Rank":31,"Models":"UniVec-CoT-7B","Model Size(B)":8.29,"Date":"unknown","Overall":
|
| 32 |
-
{"Rank":32,"Models":"OEmbedding-v1-7B","Model Size(B)":8.29,"Date":"2025-10-14","Overall":
|
| 33 |
-
{"Rank":33,"Models":"ReCo-7B","Model Size(B)":8.29,"Date":"2025-08-15","Overall":34.
|
| 34 |
-
{"Rank":34,"Models":"UniVec-7B","Model Size(B)":8.29,"Date":"unknown","Overall":34.
|
| 35 |
-
{"Rank":35,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-7B<\/a>","Model Size(B)":8.29,"Date":"2025-11-24","Overall":34.
|
| 36 |
-
{"Rank":36,"Models":"dp-embedding-v3-lite","Model Size(B)":8.29,"Date":"2025-12-15","Overall":
|
| 37 |
-
{"Rank":37,"Models":"TCE-v1","Model Size(B)":8.0,"Date":"2025-10-31","Overall":
|
| 38 |
-
{"Rank":38,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-3B<\/a>","Model Size(B)":3.75,"Date":"2025-11-24","Overall":
|
| 39 |
-
{"Rank":39,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-2B<\/a>","Model Size(B)":2.21,"Date":"2025-11-24","Overall":
|
| 40 |
-
{"Rank":40,"Models":"Taichu-UniRetriever-8B","Model Size(B)":8.77,"Date":"2026-01-15","Overall":
|
| 41 |
-
{"Rank":41,"Models":"Taichu-UniRetriever-v1-2B","Model Size(B)":2.13,"Date":"2026-01-06","Overall":
|
| 42 |
{"Rank":42,"Models":"TCR","Model Size(B)":8.77,"Date":"2026-01-15","Overall":0.0,"Image-Overall":0.0,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
|
|
|
| 1 |
+
{"Rank":1,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-8B\">Qwen3-VL-Embedding-8B<\/a>","Model Size(B)":8.14,"Date":"2026-01-06","Overall":79.87,"Image-Overall":80.12,"Video-Overall":67.15,"Visdoc-Overall":83.09}
|
| 2 |
+
{"Rank":2,"Models":"<a href=\"https:\/\/console.volcengine.com\/ark\/region:ark+cn-beijing\/model\/detail?Id=doubao-embedding-vision\">seed1.6-embedding-1215<\/a>","Model Size(B)":"unknown","Date":"2025-12-18","Overall":77.58,"Image-Overall":77.99,"Video-Overall":67.74,"Visdoc-Overall":83.16}
|
| 3 |
+
{"Rank":3,"Models":"WeMM-Embedding-8B","Model Size(B)":8.77,"Date":"2025-12-16","Overall":76.38,"Image-Overall":78.09,"Video-Overall":63.24,"Visdoc-Overall":82.49}
|
| 4 |
+
{"Rank":4,"Models":"<a href=\"https:\/\/interestfm-tte.github.io\/\">IFM-TTE-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-11","Overall":76.02,"Image-Overall":77.9,"Video-Overall":59.19,"Visdoc-Overall":83.14}
|
| 5 |
+
{"Rank":5,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-2B\">Qwen3-VL-Embedding-2B<\/a>","Model Size(B)":2.13,"Date":"2026-01-06","Overall":75.18,"Image-Overall":74.96,"Video-Overall":61.87,"Visdoc-Overall":80.16}
|
| 6 |
+
{"Rank":6,"Models":"WeMM-Embedding-2B","Model Size(B)":2.13,"Date":"2025-12-16","Overall":73.56,"Image-Overall":76.08,"Video-Overall":58.67,"Visdoc-Overall":79.85}
|
| 7 |
+
{"Rank":7,"Models":"<a href=\"https:\/\/github.com\/360CVGroup\/RzenEmbed\">RzenEmbed-v2-7B<\/a>","Model Size(B)":8.29,"Date":"2025-09-20","Overall":73.5,"Image-Overall":75.92,"Video-Overall":55.73,"Visdoc-Overall":82.31}
|
| 8 |
+
{"Rank":8,"Models":"<a href=\"https:\/\/seed1-6-embedding.github.io\">seed-1.6-embedding<\/a>","Model Size(B)":"unknown","Date":"2025-06-17","Overall":73.14,"Image-Overall":77.78,"Video-Overall":55.34,"Visdoc-Overall":78.37}
|
| 9 |
+
{"Rank":9,"Models":"RzenEmbed-v1-7B","Model Size(B)":8.29,"Date":"2025-07-30","Overall":70.69,"Image-Overall":73.6,"Video-Overall":48.87,"Visdoc-Overall":82.0}
|
| 10 |
+
{"Rank":10,"Models":"<a href=\"https:\/\/huggingface.co\/OpenSearch-AI\/Ops-MM-embedding-v1-7B\">Ops-MM-embedding-v1-7B<\/a>","Model Size(B)":8.29,"Date":"2025-07-02","Overall":69.39,"Image-Overall":72.72,"Video-Overall":53.76,"Visdoc-Overall":74.99}
|
| 11 |
+
{"Rank":11,"Models":"<a href=\"https:\/\/huggingface.co\/Haon-Chen\/e5-omni-7B\">e5-omni-7B<\/a>","Model Size(B)":8.0,"Date":"2026-01-09","Overall":68.1,"Image-Overall":71.23,"Video-Overall":43.54,"Visdoc-Overall":81.3}
|
| 12 |
+
{"Rank":12,"Models":"Crotchet-embedding-2B","Model Size(B)":2.13,"Date":"2025-12-31","Overall":67.3,"Image-Overall":68.18,"Video-Overall":50.63,"Visdoc-Overall":73.59}
|
| 13 |
+
{"Rank":13,"Models":"<a href=\"https:\/\/huggingface.co\/zhibinlan\/UME-R1-7B\">UME-R1-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-20","Overall":66.2,"Image-Overall":71.25,"Video-Overall":47.5,"Visdoc-Overall":71.71}
|
| 14 |
+
{"Rank":14,"Models":"RzenEmbed-v1-2B","Model Size(B)":2.21,"Date":"2025-07-16","Overall":66.05,"Image-Overall":68.53,"Video-Overall":42.62,"Visdoc-Overall":79.4}
|
| 15 |
+
{"Rank":15,"Models":"<a href=\"https:\/\/huggingface.co\/OpenSearch-AI\/Ops-MM-embedding-v1-2B\">Ops-MM-embedding-v1-2B<\/a>","Model Size(B)":2.21,"Date":"2025-07-02","Overall":65.11,"Image-Overall":69.03,"Video-Overall":47.56,"Visdoc-Overall":71.48}
|
| 16 |
+
{"Rank":16,"Models":"<a href=\"https:\/\/huggingface.co\/Haon-Chen\/e5-omni-3B\">e5-omni-3B<\/a>","Model Size(B)":3.0,"Date":"2026-01-09","Overall":64.74,"Image-Overall":67.59,"Video-Overall":40.63,"Visdoc-Overall":78.11}
|
| 17 |
+
{"Rank":17,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-7B<\/a>","Model Size(B)":8.03,"Date":"2025-07-06","Overall":62.23,"Image-Overall":67.56,"Video-Overall":42.4,"Visdoc-Overall":68.1}
|
| 18 |
+
{"Rank":18,"Models":"<a href=\"https:\/\/huggingface.co\/zhibinlan\/UME-R1-2B\">UME-R1-2B<\/a>","Model Size(B)":2.21,"Date":"2025-10-20","Overall":61.69,"Image-Overall":66.56,"Video-Overall":42.23,"Visdoc-Overall":68.16}
|
| 19 |
+
{"Rank":19,"Models":"<a href=\"https:\/\/github.com\/GaryGuTC\/UniME-v2\">UniME-V2-LLaVA-OneVision-7B<\/a>","Model Size(B)":8.03,"Date":"2025-10-15","Overall":61.13,"Image-Overall":71.77,"Video-Overall":39.01,"Visdoc-Overall":60.24}
|
| 20 |
+
{"Rank":20,"Models":"<a href=\"https:\/\/huggingface.co\/VLM2Vec\/VLM2Vec-V2.0\">VLM2Vec-V2.0-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Overall":59.01,"Image-Overall":64.92,"Video-Overall":34.67,"Visdoc-Overall":63.56}
|
| 21 |
+
{"Rank":21,"Models":"<a href=\"https:\/\/huggingface.co\/Alibaba-NLP\/gme-Qwen2-VL-7B-Instruct\">gme-Qwen2-VL-7B-Instruct<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":58.83,"Image-Overall":55.93,"Video-Overall":38.43,"Visdoc-Overall":73.96}
|
| 22 |
+
{"Rank":22,"Models":"<a href=\"https:\/\/huggingface.co\/Alibaba-NLP\/gme-Qwen2-VL-2B-Instruct\">gme-Qwen2-VL-2B-Instruct<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Overall":55.06,"Image-Overall":51.87,"Video-Overall":33.59,"Visdoc-Overall":71.88}
|
| 23 |
+
{"Rank":23,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-0.5B<\/a>","Model Size(B)":0.894,"Date":"2025-07-06","Overall":50.99,"Image-Overall":55.43,"Video-Overall":35.87,"Visdoc-Overall":54.83}
|
| 24 |
+
{"Rank":24,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-7B\">VLM2Vec-V1-Qwen2VL-7B<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":48.23,"Image-Overall":65.37,"Video-Overall":33.79,"Visdoc-Overall":26.82}
|
| 25 |
+
{"Rank":25,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret-Qwen2.5VL-7b\">LamRA-Ret-Qwen2.5VL-7b<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":48.05,"Image-Overall":52.54,"Video-Overall":33.83,"Visdoc-Overall":47.61}
|
| 26 |
+
{"Rank":26,"Models":"colpali-v1.3","Model Size(B)":"unknown","Date":"2025-12-29","Overall":45.36,"Image-Overall":34.89,"Video-Overall":28.23,"Visdoc-Overall":70.5}
|
| 27 |
+
{"Rank":27,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-2B\">VLM2Vec-V1-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Overall":43.81,"Image-Overall":59.73,"Video-Overall":28.55,"Visdoc-Overall":26.0}
|
| 28 |
+
{"Rank":28,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret\">LamRA-Ret<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Overall":42.7,"Image-Overall":54.13,"Video-Overall":35.21,"Visdoc-Overall":27.72}
|
| 29 |
+
{"Rank":29,"Models":"QQMM-embed-v3","Model Size(B)":8.29,"Date":"2025-12-30","Overall":37.05,"Image-Overall":78.21,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 30 |
+
{"Rank":30,"Models":"<a href=\"https:\/\/github.com\/QQ-MM\/QQMM-embed\">QQMM-embed-v2<\/a>","Model Size(B)":8.29,"Date":"2025-09-15","Overall":35.66,"Image-Overall":75.28,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 31 |
+
{"Rank":31,"Models":"UniVec-CoT-7B","Model Size(B)":8.29,"Date":"unknown","Overall":35.1,"Image-Overall":74.11,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 32 |
+
{"Rank":32,"Models":"OEmbedding-v1-7B","Model Size(B)":8.29,"Date":"2025-10-14","Overall":35.08,"Image-Overall":74.05,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 33 |
+
{"Rank":33,"Models":"ReCo-7B","Model Size(B)":8.29,"Date":"2025-08-15","Overall":34.99,"Image-Overall":73.87,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 34 |
+
{"Rank":34,"Models":"UniVec-7B","Model Size(B)":8.29,"Date":"unknown","Overall":34.96,"Image-Overall":73.79,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 35 |
+
{"Rank":35,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-7B<\/a>","Model Size(B)":8.29,"Date":"2025-11-24","Overall":34.91,"Image-Overall":73.7,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 36 |
+
{"Rank":36,"Models":"dp-embedding-v3-lite","Model Size(B)":8.29,"Date":"2025-12-15","Overall":34.76,"Image-Overall":73.38,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 37 |
+
{"Rank":37,"Models":"TCE-v1","Model Size(B)":8.0,"Date":"2025-10-31","Overall":34.27,"Image-Overall":72.36,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 38 |
+
{"Rank":38,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-3B<\/a>","Model Size(B)":3.75,"Date":"2025-11-24","Overall":33.25,"Image-Overall":70.2,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 39 |
+
{"Rank":39,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-2B<\/a>","Model Size(B)":2.21,"Date":"2025-11-24","Overall":32.78,"Image-Overall":69.21,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
| 40 |
+
{"Rank":40,"Models":"Taichu-UniRetriever-8B","Model Size(B)":8.77,"Date":"2026-01-15","Overall":26.29,"Image-Overall":0.0,"Video-Overall":0.0,"Visdoc-Overall":83.72}
|
| 41 |
+
{"Rank":41,"Models":"Taichu-UniRetriever-v1-2B","Model Size(B)":2.13,"Date":"2026-01-06","Overall":24.21,"Image-Overall":0.0,"Video-Overall":0.0,"Visdoc-Overall":81.93}
|
| 42 |
{"Rank":42,"Models":"TCR","Model Size(B)":8.77,"Date":"2026-01-15","Overall":0.0,"Image-Overall":0.0,"Video-Overall":0.0,"Visdoc-Overall":0.0}
|
rankings/visdoc_ranking.csv
CHANGED
|
@@ -1,43 +1,43 @@
|
|
| 1 |
-
Rank,Models,Model Size(B),Date,Visdoc-Overall,ViDoRe-V1,ViDoRe-V2,VisRAG,VisDoc-OOD,ViDoRe_arxivqa,ViDoRe_docvqa,ViDoRe_infovqa,ViDoRe_tabfquad,ViDoRe_tatdqa,ViDoRe_shiftproject,ViDoRe_syntheticDocQA_artificial_intelligence,ViDoRe_syntheticDocQA_energy,ViDoRe_syntheticDocQA_government_reports,ViDoRe_syntheticDocQA_healthcare_industry,ViDoRe_esg_reports_human_labeled_v2,ViDoRe_biomedical_lectures_v2_multilingual,ViDoRe_economics_reports_v2_multilingual,ViDoRe_esg_reports_v2_multilingual,VisRAG_ArxivQA,VisRAG_ChartQA,VisRAG_MP-DocVQA,VisRAG_SlideVQA,VisRAG_InfoVQA,VisRAG_PlotQA,ViDoSeek-
|
| 2 |
-
1,Taichu-UniRetriever-8B,8.77,2026-01-15,83.
|
| 3 |
-
2,"<a href=""https://
|
| 4 |
-
3,"<a href=""https://interestfm-tte.github.io/"">IFM-TTE-7B</a>",8.29,2025-10-11,
|
| 5 |
-
4,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-
|
| 6 |
-
5,
|
| 7 |
-
6,
|
| 8 |
-
7,
|
| 9 |
-
8,
|
| 10 |
-
9,
|
| 11 |
-
10,"<a href=""https://huggingface.co/
|
| 12 |
-
11,WeMM-Embedding-2B,2.13,2025-12-16,
|
| 13 |
-
12,RzenEmbed-v1-2B,2.21,2025-07-16,
|
| 14 |
-
13,"<a href=""https://
|
| 15 |
-
14,"<a href=""https://
|
| 16 |
-
15,"<a href=""https://huggingface.co/
|
| 17 |
-
16,
|
| 18 |
-
17,
|
| 19 |
-
18,"<a href=""https://huggingface.co/
|
| 20 |
-
19,
|
| 21 |
-
20,"<a href=""https://huggingface.co/
|
| 22 |
-
21,
|
| 23 |
-
22,"<a href=""https://
|
| 24 |
-
23,"<a href=""https://
|
| 25 |
-
24,"<a href=""https://huggingface.co/VLM2Vec/VLM2Vec-V2.0"">VLM2Vec-V2.0-Qwen2VL-2B</a>",2.21,2025-12-29,63.
|
| 26 |
-
25,"<a href=""https://github.com/GaryGuTC/UniME-v2"">UniME-V2-LLaVA-OneVision-7B</a>",8.03,2025-10-15,
|
| 27 |
-
26,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-0.5B</a>",0.894,2025-07-06,
|
| 28 |
-
27,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret-Qwen2.5VL-7b"">LamRA-Ret-Qwen2.5VL-7b</a>",8.29,2025-12-29,47.
|
| 29 |
-
28,"<a href=""https://huggingface.co/
|
| 30 |
-
29,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-
|
| 31 |
-
30,"<a href=""https://huggingface.co/
|
| 32 |
-
31,dp-embedding-v3-lite,8.29,2025-12-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 33 |
-
32,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-2B</a>",2.21,2025-11-24,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 34 |
-
33,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-3B</a>",3.75,2025-11-24,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 35 |
-
34,TCE-v1,8.0,2025-10-31,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 36 |
-
35,
|
| 37 |
-
36,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-7B</a>",8.29,2025-11-24,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 38 |
-
37,UniVec-7B,8.29,unknown,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 39 |
-
38,ReCo-7B,8.29,2025-08-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 40 |
-
39,OEmbedding-v1-7B,8.29,2025-10-14,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 41 |
-
40,
|
| 42 |
-
41,QQMM-embed-v3,8.29,2025-12-30,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
| 43 |
-
42,TCR,8.77,2026-01-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,
|
|
|
|
| 1 |
+
Rank,Models,Model Size(B),Date,Visdoc-Overall,ViDoRe-V1,ViDoRe-V2,VisRAG,VisDoc-OOD,ViDoRe_arxivqa,ViDoRe_docvqa,ViDoRe_infovqa,ViDoRe_tabfquad,ViDoRe_tatdqa,ViDoRe_shiftproject,ViDoRe_syntheticDocQA_artificial_intelligence,ViDoRe_syntheticDocQA_energy,ViDoRe_syntheticDocQA_government_reports,ViDoRe_syntheticDocQA_healthcare_industry,ViDoRe_esg_reports_human_labeled_v2,ViDoRe_biomedical_lectures_v2_multilingual,ViDoRe_economics_reports_v2_multilingual,ViDoRe_esg_reports_v2_multilingual,VisRAG_ArxivQA,VisRAG_ChartQA,VisRAG_MP-DocVQA,VisRAG_SlideVQA,VisRAG_InfoVQA,VisRAG_PlotQA,ViDoSeek-doc,MMLongBench-doc,ViDoSeek-page-fixed,MMLongBench-page-fixed,ViDoSeek-page-old,MMLongBench-page-old
|
| 2 |
+
1,Taichu-UniRetriever-8B,8.77,2026-01-15,83.72,90.93,62.91,89.66,71.45,90.06,61.9,93.93,97.33,80.53,93.2,98.89,97.5,97.05,98.89,69.67,65.62,55.32,61.02,88.58,90.2,91.47,97.39,97.03,73.3,85.62,57.28,92.42,63.78,✅,✅
|
| 3 |
+
2,"<a href=""https://console.volcengine.com/ark/region:ark+cn-beijing/model/detail?Id=doubao-embedding-vision"">seed1.6-embedding-1215</a>",unknown,2025-12-18,83.16,90.9,60.31,89.96,69.82,92.29,61.84,94.28,94.99,80.41,91.71,99.26,96.76,98.15,99.26,70.52,63.46,55.44,51.82,90.49,87.59,94.07,97.3,95.87,74.47,84.27,55.37,⚠️! Please fix this score!,⚠️! Please fix this score!,22.29,17.15
|
| 4 |
+
3,"<a href=""https://interestfm-tte.github.io/"">IFM-TTE-7B</a>",8.29,2025-10-11,83.14,85.19,71.5,92.75,67.31,84.53,45.76,88.33,94.28,55.31,89.76,99.26,97.26,98.28,99.13,87.5,64.85,53.94,79.73,92.73,95.07,87.31,95.9,93.0,92.49,82.6,52.03,⚠️! Please fix this score!,⚠️! Please fix this score!,50.29,28.18
|
| 5 |
+
4,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-8B"">Qwen3-VL-Embedding-8B</a>",8.14,2026-01-06,83.09,87.21,69.86,88.68,72.19,87.0,54.02,90.86,96.67,69.97,84.14,99.26,94.41,98.02,97.79,71.39,71.59,67.18,69.28,88.21,88.34,89.27,96.73,94.75,74.8,86.1,58.29,88.38,60.29,✅,✅
|
| 6 |
+
5,WeMM-Embedding-8B,8.77,2025-12-16,82.49,89.47,59.31,90.43,70.13,92.22,60.95,92.83,93.41,77.03,88.5,98.15,96.21,96.17,99.26,62.97,61.9,56.85,55.51,90.42,91.4,91.82,97.0,95.97,75.95,84.31,55.95,⚠️! Please fix this score!,⚠️! Please fix this score!,23.34,17.05
|
| 7 |
+
6,"<a href=""https://github.com/360CVGroup/RzenEmbed"">RzenEmbed-v2-7B</a>",8.29,2025-09-20,82.31,89.7,60.7,88.7,69.4,88.42,59.61,92.85,95.85,76.5,91.76,99.63,95.71,97.02,99.63,65.61,61.71,59.1,56.39,89.38,89.61,90.59,95.68,94.11,72.81,84.02,54.78,⚠️! Please fix this score!,⚠️! Please fix this score!,23.09,15.62
|
| 8 |
+
7,RzenEmbed-v1-7B,8.29,2025-07-30,82.0,89.47,60.77,87.92,69.3,86.89,57.99,92.24,96.25,75.22,93.35,99.63,95.25,98.28,99.63,63.67,61.58,60.8,57.05,86.97,88.34,90.7,95.75,94.37,71.39,83.92,54.67,⚠️! Please fix this score!,⚠️! Please fix this score!,23.06,16.1
|
| 9 |
+
8,Taichu-UniRetriever-v1-2B,2.13,2026-01-06,81.93,89.81,57.34,88.79,71.13,88.14,59.09,92.63,96.67,80.07,89.93,99.26,98.06,95.93,98.32,63.6,62.47,51.21,52.08,87.25,87.33,91.99,96.93,95.9,73.32,85.07,57.19,⚠️! Please fix this score!,⚠️! Please fix this score!,20.92,16.49
|
| 10 |
+
9,"<a href=""https://huggingface.co/Haon-Chen/e5-omni-7B"">e5-omni-7B</a>",8.0,2026-01-09,81.3,87.58,62.39,87.46,69.21,87.57,57.96,92.29,93.19,70.9,85.79,98.15,94.95,96.28,98.69,64.03,64.48,60.34,60.71,88.02,89.02,89.07,95.99,93.54,69.15,83.44,54.98,⚠️! Please fix this score!,⚠️! Please fix this score!,23.11,15.78
|
| 11 |
+
10,"<a href=""https://huggingface.co/Qwen/Qwen3-VL-Embedding-2B"">Qwen3-VL-Embedding-2B</a>",2.13,2026-01-06,80.16,84.37,65.34,86.43,70.0,81.8,49.36,88.35,96.4,63.19,83.9,97.05,90.84,95.68,97.09,64.56,68.23,63.09,65.49,83.58,87.13,87.06,95.59,92.88,72.33,85.0,54.99,81.94,55.73,✅,✅
|
| 12 |
+
11,WeMM-Embedding-2B,2.13,2025-12-16,79.85,87.22,53.4,88.84,68.89,90.25,59.11,91.2,89.41,73.81,82.61,98.15,94.08,95.35,98.28,54.54,55.64,53.93,49.5,88.62,88.61,91.01,95.83,93.74,75.23,83.14,54.65,⚠️! Please fix this score!,⚠️! Please fix this score!,22.27,16.53
|
| 13 |
+
12,RzenEmbed-v1-2B,2.21,2025-07-16,79.4,86.98,57.63,85.35,67.11,83.4,56.09,89.89,94.39,69.32,89.88,98.76,92.78,97.19,98.15,58.93,59.87,57.98,53.73,83.68,86.39,85.54,94.07,91.92,70.52,82.27,51.95,⚠️! Please fix this score!,⚠️! Please fix this score!,22.89,16.21
|
| 14 |
+
13,"<a href=""https://seed1-6-embedding.github.io"">seed-1.6-embedding</a>",unknown,2025-06-17,78.37,85.53,56.57,84.74,67.09,88.46,60.36,90.62,87.82,74.98,73.34,96.32,92.83,93.54,97.05,63.3,57.14,53.85,51.99,85.79,81.88,89.43,94.9,92.41,64.02,82.57,51.61,⚠️! Please fix this score!,⚠️! Please fix this score!,22.8,15.57
|
| 15 |
+
14,"<a href=""https://huggingface.co/Haon-Chen/e5-omni-3B"">e5-omni-3B</a>",3.0,2026-01-09,78.11,83.13,59.53,85.48,68.07,82.2,51.99,88.88,88.86,58.78,80.14,98.15,93.17,93.21,95.92,61.84,60.92,64.39,50.97,84.52,86.75,85.51,95.4,92.55,68.17,83.29,52.85,⚠️! Please fix this score!,⚠️! Please fix this score!,22.38,15.0
|
| 16 |
+
15,"<a href=""https://huggingface.co/OpenSearch-AI/Ops-MM-embedding-v1-7B"">Ops-MM-embedding-v1-7B</a>",8.29,2025-07-02,74.99,80.05,59.59,79.31,67.49,78.17,49.11,86.59,91.4,55.59,76.17,90.99,87.53,91.56,93.39,66.27,54.34,60.92,56.82,78.18,79.53,78.2,91.11,87.16,61.71,83.47,51.51,⚠️! Please fix this score!,⚠️! Please fix this score!,22.47,15.89
|
| 17 |
+
16,"<a href=""https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-7B-Instruct"">gme-Qwen2-VL-7B-Instruct</a>",8.29,2025-12-29,73.96,89.55,55.47,84.95,0.0,87.53,56.59,92.17,92.71,76.58,95.64,99.63,95.69,99.5,99.5,62.76,49.78,53.91,55.43,87.69,81.29,89.08,94.72,93.49,63.44,0.0,0.0,83.57,55.32,✅,✅
|
| 18 |
+
17,Crotchet-embedding-2B,2.13,2025-12-31,73.59,78.16,53.28,82.21,65.51,84.34,42.93,85.18,93.29,53.04,71.02,90.74,85.07,85.45,90.59,62.65,46.92,51.47,52.09,81.64,85.47,76.0,90.85,88.54,70.74,81.84,49.18,⚠️! Please fix this score!,⚠️! Please fix this score!,78.36,51.39
|
| 19 |
+
18,"<a href=""https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-2B-Instruct"">gme-Qwen2-VL-2B-Instruct</a>",2.21,2025-12-29,71.88,87.12,53.88,82.43,0.0,82.53,55.16,90.69,93.31,70.26,92.85,98.13,92.56,97.15,98.52,60.29,53.99,50.8,50.43,81.95,79.09,84.34,93.68,91.41,64.12,0.0,0.0,79.81,51.32,✅,✅
|
| 20 |
+
19,"<a href=""https://huggingface.co/zhibinlan/UME-R1-7B"">UME-R1-7B</a>",8.29,2025-10-20,71.71,75.66,50.53,83.7,58.3,73.56,41.1,80.83,90.24,46.67,64.96,89.47,85.7,89.82,94.29,50.41,50.73,57.8,43.18,80.45,84.95,83.43,91.47,89.2,72.7,75.33,41.28,⚠️! Please fix this score!,⚠️! Please fix this score!,21.28,12.32
|
| 21 |
+
20,"<a href=""https://huggingface.co/OpenSearch-AI/Ops-MM-embedding-v1-2B"">Ops-MM-embedding-v1-2B</a>",2.21,2025-07-02,71.48,76.39,53.18,77.64,65.09,73.72,45.01,81.41,88.81,49.59,72.57,89.8,84.27,87.06,91.64,58.57,52.87,47.89,53.39,73.84,80.17,75.06,89.15,86.66,60.99,82.29,47.89,⚠️! Please fix this score!,⚠️! Please fix this score!,21.44,13.06
|
| 22 |
+
21,colpali-v1.3,unknown,2025-12-29,70.5,84.57,54.81,81.0,0.0,83.33,59.2,85.38,86.63,70.47,76.83,96.26,95.02,95.59,97.02,57.25,56.25,50.2,55.56,80.9,77.04,87.03,95.13,85.73,60.18,0.0,0.0,79.22,53.18,✅,✅
|
| 23 |
+
22,"<a href=""https://huggingface.co/zhibinlan/UME-R1-2B"">UME-R1-2B</a>",2.21,2025-10-20,68.16,72.39,46.16,79.22,57.78,73.86,37.85,76.19,86.11,40.63,66.79,85.87,83.25,82.56,90.83,50.16,46.15,45.69,42.65,74.26,85.98,75.59,87.11,84.38,67.98,75.89,39.68,⚠️! Please fix this score!,⚠️! Please fix this score!,21.2,11.9
|
| 24 |
+
23,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-7B</a>",8.03,2025-07-06,68.1,70.68,49.57,79.45,58.21,73.29,38.27,80.59,80.7,37.79,52.02,85.99,84.78,84.96,88.37,50.67,50.89,54.38,42.33,73.96,82.71,75.15,87.58,87.91,69.42,73.82,42.61,⚠️! Please fix this score!,⚠️! Please fix this score!,22.52,13.32
|
| 25 |
+
24,"<a href=""https://huggingface.co/VLM2Vec/VLM2Vec-V2.0"">VLM2Vec-V2.0-Qwen2VL-2B</a>",2.21,2025-12-29,63.56,74.39,44.61,79.33,0.0,78.87,37.06,82.71,87.77,44.29,61.04,89.12,86.29,85.58,91.13,45.79,44.59,42.33,45.74,76.74,84.2,71.8,91.44,85.88,65.91,0.0,0.0,80.3,44.7,✅,✅
|
| 26 |
+
25,"<a href=""https://github.com/GaryGuTC/UniME-v2"">UniME-V2-LLaVA-OneVision-7B</a>",8.03,2025-10-15,60.24,61.76,42.0,70.53,58.3,51.9,38.17,73.16,57.75,35.48,45.42,76.85,77.27,79.87,81.71,54.67,33.75,35.95,43.63,53.13,83.67,66.4,86.44,82.58,50.96,75.79,40.8,⚠️! Please fix this score!,⚠️! Please fix this score!,22.89,11.96
|
| 27 |
+
26,"<a href=""https://arxiv.org/abs/2503.19900"">interestFM-UIR-CAFe-0.5B</a>",0.894,2025-07-06,54.83,56.93,32.59,68.57,47.59,51.19,25.46,71.95,59.64,27.05,31.8,78.74,70.75,75.47,77.29,36.24,29.72,37.61,26.78,57.82,75.62,63.24,82.32,80.28,52.14,61.41,33.78,⚠️! Please fix this score!,⚠️! Please fix this score!,17.62,9.95
|
| 28 |
+
27,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret-Qwen2.5VL-7b"">LamRA-Ret-Qwen2.5VL-7b</a>",8.29,2025-12-29,47.61,56.4,33.62,58.15,0.0,53.04,25.42,72.19,66.51,26.01,28.13,70.31,66.33,72.21,83.8,32.98,35.87,32.37,33.26,38.03,65.21,54.14,76.55,73.5,41.46,0.0,0.0,68.54,35.83,✅,✅
|
| 29 |
+
28,"<a href=""https://huggingface.co/code-kunkun/LamRA-Ret"">LamRA-Ret</a>",8.29,2025-12-29,27.72,33.83,11.53,37.56,0.0,31.54,19.85,63.67,53.47,7.88,15.95,29.83,36.05,41.18,38.83,6.91,13.42,19.37,6.43,2.04,42.74,33.37,56.3,56.86,34.06,0.0,0.0,34.48,18.29,✅,✅
|
| 30 |
+
29,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-7B"">VLM2Vec-V1-Qwen2VL-7B</a>",8.29,2025-12-29,26.82,19.99,9.24,58.85,0.0,28.16,18.98,44.77,16.98,5.73,1.63,18.23,23.9,13.88,27.61,6.95,5.22,13.75,11.05,52.85,68.96,52.74,72.76,71.3,34.52,0.0,0.0,77.4,36.76,✅,✅
|
| 31 |
+
30,"<a href=""https://huggingface.co/TIGER-Lab/VLM2Vec-Qwen2VL-2B"">VLM2Vec-V1-Qwen2VL-2B</a>",2.21,2025-12-29,26.0,20.59,13.19,52.23,0.0,18.12,14.01,39.53,36.03,10.49,8.41,17.04,16.35,25.15,20.79,13.05,6.51,12.86,20.32,41.22,59.48,43.55,74.5,71.1,23.51,0.0,0.0,67.81,26.04,✅,✅
|
| 32 |
+
31,dp-embedding-v3-lite,8.29,2025-12-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 33 |
+
32,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-2B</a>",2.21,2025-11-24,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 34 |
+
33,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-3B</a>",3.75,2025-11-24,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 35 |
+
34,TCE-v1,8.0,2025-10-31,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 36 |
+
35,UniVec-CoT-7B,8.29,unknown,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 37 |
+
36,"<a href=""https://arxiv.org/abs/2511.19278"">ReMatch-7B</a>",8.29,2025-11-24,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 38 |
+
37,UniVec-7B,8.29,unknown,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 39 |
+
38,ReCo-7B,8.29,2025-08-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 40 |
+
39,OEmbedding-v1-7B,8.29,2025-10-14,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 41 |
+
40,"<a href=""https://github.com/QQ-MM/QQMM-embed"">QQMM-embed-v2</a>",8.29,2025-09-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 42 |
+
41,QQMM-embed-v3,8.29,2025-12-30,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
| 43 |
+
42,TCR,8.77,2026-01-15,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,✅,✅
|
rankings/visdoc_ranking.jsonl
CHANGED
|
@@ -1,42 +1,42 @@
|
|
| 1 |
-
{"Rank":1,"Models":"Taichu-UniRetriever-8B","Model Size(B)":8.77,"Date":"2026-01-15","Visdoc-Overall":83.
|
| 2 |
-
{"Rank":2,"Models":"<a href=\"https:\/\/
|
| 3 |
-
{"Rank":3,"Models":"<a href=\"https:\/\/interestfm-tte.github.io\/\">IFM-TTE-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-11","Visdoc-Overall":
|
| 4 |
-
{"Rank":4,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-
|
| 5 |
-
{"Rank":5,"Models":"
|
| 6 |
-
{"Rank":6,"Models":"
|
| 7 |
-
{"Rank":7,"Models":"
|
| 8 |
-
{"Rank":8,"Models":"
|
| 9 |
-
{"Rank":9,"Models":"
|
| 10 |
-
{"Rank":10,"Models":"<a href=\"https:\/\/huggingface.co\/
|
| 11 |
-
{"Rank":11,"Models":"WeMM-Embedding-2B","Model Size(B)":2.13,"Date":"2025-12-16","Visdoc-Overall":
|
| 12 |
-
{"Rank":12,"Models":"RzenEmbed-v1-2B","Model Size(B)":2.21,"Date":"2025-07-16","Visdoc-Overall":
|
| 13 |
-
{"Rank":13,"Models":"<a href=\"https:\/\/
|
| 14 |
-
{"Rank":14,"Models":"<a href=\"https:\/\/
|
| 15 |
-
{"Rank":15,"Models":"<a href=\"https:\/\/huggingface.co\/
|
| 16 |
-
{"Rank":16,"Models":"
|
| 17 |
-
{"Rank":17,"Models":"
|
| 18 |
-
{"Rank":18,"Models":"<a href=\"https:\/\/huggingface.co\/
|
| 19 |
-
{"Rank":19,"Models":"
|
| 20 |
-
{"Rank":20,"Models":"<a href=\"https:\/\/huggingface.co\/
|
| 21 |
-
{"Rank":21,"Models":"
|
| 22 |
-
{"Rank":22,"Models":"<a href=\"https:\/\/
|
| 23 |
-
{"Rank":23,"Models":"<a href=\"https:\/\/
|
| 24 |
-
{"Rank":24,"Models":"<a href=\"https:\/\/huggingface.co\/VLM2Vec\/VLM2Vec-V2.0\">VLM2Vec-V2.0-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Visdoc-Overall":63.
|
| 25 |
-
{"Rank":25,"Models":"<a href=\"https:\/\/github.com\/GaryGuTC\/UniME-v2\">UniME-V2-LLaVA-OneVision-7B<\/a>","Model Size(B)":8.03,"Date":"2025-10-15","Visdoc-Overall":
|
| 26 |
-
{"Rank":26,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-0.5B<\/a>","Model Size(B)":0.894,"Date":"2025-07-06","Visdoc-Overall":
|
| 27 |
-
{"Rank":27,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret-Qwen2.5VL-7b\">LamRA-Ret-Qwen2.5VL-7b<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Visdoc-Overall":47.
|
| 28 |
-
{"Rank":28,"Models":"<a href=\"https:\/\/huggingface.co\/
|
| 29 |
-
{"Rank":29,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-
|
| 30 |
-
{"Rank":30,"Models":"<a href=\"https:\/\/huggingface.co\/
|
| 31 |
-
{"Rank":31,"Models":"dp-embedding-v3-lite","Model Size(B)":8.29,"Date":"2025-12-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 32 |
-
{"Rank":32,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-2B<\/a>","Model Size(B)":2.21,"Date":"2025-11-24","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 33 |
-
{"Rank":33,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-3B<\/a>","Model Size(B)":3.75,"Date":"2025-11-24","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 34 |
-
{"Rank":34,"Models":"TCE-v1","Model Size(B)":8.0,"Date":"2025-10-31","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 35 |
-
{"Rank":35,"Models":"
|
| 36 |
-
{"Rank":36,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-7B<\/a>","Model Size(B)":8.29,"Date":"2025-11-24","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 37 |
-
{"Rank":37,"Models":"UniVec-7B","Model Size(B)":8.29,"Date":"unknown","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 38 |
-
{"Rank":38,"Models":"ReCo-7B","Model Size(B)":8.29,"Date":"2025-08-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 39 |
-
{"Rank":39,"Models":"OEmbedding-v1-7B","Model Size(B)":8.29,"Date":"2025-10-14","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 40 |
-
{"Rank":40,"Models":"
|
| 41 |
-
{"Rank":41,"Models":"QQMM-embed-v3","Model Size(B)":8.29,"Date":"2025-12-30","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
| 42 |
-
{"Rank":42,"Models":"TCR","Model Size(B)":8.77,"Date":"2026-01-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-
|
|
|
|
| 1 |
+
{"Rank":1,"Models":"Taichu-UniRetriever-8B","Model Size(B)":8.77,"Date":"2026-01-15","Visdoc-Overall":83.72,"ViDoRe-V1":90.93,"ViDoRe-V2":62.91,"VisRAG":89.66,"VisDoc-OOD":71.45,"ViDoRe_arxivqa":90.06,"ViDoRe_docvqa":61.9,"ViDoRe_infovqa":93.93,"ViDoRe_tabfquad":97.33,"ViDoRe_tatdqa":80.53,"ViDoRe_shiftproject":93.2,"ViDoRe_syntheticDocQA_artificial_intelligence":98.89,"ViDoRe_syntheticDocQA_energy":97.5,"ViDoRe_syntheticDocQA_government_reports":97.05,"ViDoRe_syntheticDocQA_healthcare_industry":98.89,"ViDoRe_esg_reports_human_labeled_v2":69.67,"ViDoRe_biomedical_lectures_v2_multilingual":65.62,"ViDoRe_economics_reports_v2_multilingual":55.32,"ViDoRe_esg_reports_v2_multilingual":61.02,"VisRAG_ArxivQA":88.58,"VisRAG_ChartQA":90.2,"VisRAG_MP-DocVQA":91.47,"VisRAG_SlideVQA":97.39,"VisRAG_InfoVQA":97.03,"VisRAG_PlotQA":73.3,"ViDoSeek-doc":85.62,"MMLongBench-doc":57.28,"ViDoSeek-page-fixed":92.42,"MMLongBench-page-fixed":63.78,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 2 |
+
{"Rank":2,"Models":"<a href=\"https:\/\/console.volcengine.com\/ark\/region:ark+cn-beijing\/model\/detail?Id=doubao-embedding-vision\">seed1.6-embedding-1215<\/a>","Model Size(B)":"unknown","Date":"2025-12-18","Visdoc-Overall":83.16,"ViDoRe-V1":90.9,"ViDoRe-V2":60.31,"VisRAG":89.96,"VisDoc-OOD":69.82,"ViDoRe_arxivqa":92.29,"ViDoRe_docvqa":61.84,"ViDoRe_infovqa":94.28,"ViDoRe_tabfquad":94.99,"ViDoRe_tatdqa":80.41,"ViDoRe_shiftproject":91.71,"ViDoRe_syntheticDocQA_artificial_intelligence":99.26,"ViDoRe_syntheticDocQA_energy":96.76,"ViDoRe_syntheticDocQA_government_reports":98.15,"ViDoRe_syntheticDocQA_healthcare_industry":99.26,"ViDoRe_esg_reports_human_labeled_v2":70.52,"ViDoRe_biomedical_lectures_v2_multilingual":63.46,"ViDoRe_economics_reports_v2_multilingual":55.44,"ViDoRe_esg_reports_v2_multilingual":51.82,"VisRAG_ArxivQA":90.49,"VisRAG_ChartQA":87.59,"VisRAG_MP-DocVQA":94.07,"VisRAG_SlideVQA":97.3,"VisRAG_InfoVQA":95.87,"VisRAG_PlotQA":74.47,"ViDoSeek-doc":84.27,"MMLongBench-doc":55.37,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.29,"MMLongBench-page-old":17.15}
|
| 3 |
+
{"Rank":3,"Models":"<a href=\"https:\/\/interestfm-tte.github.io\/\">IFM-TTE-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-11","Visdoc-Overall":83.14,"ViDoRe-V1":85.19,"ViDoRe-V2":71.5,"VisRAG":92.75,"VisDoc-OOD":67.31,"ViDoRe_arxivqa":84.53,"ViDoRe_docvqa":45.76,"ViDoRe_infovqa":88.33,"ViDoRe_tabfquad":94.28,"ViDoRe_tatdqa":55.31,"ViDoRe_shiftproject":89.76,"ViDoRe_syntheticDocQA_artificial_intelligence":99.26,"ViDoRe_syntheticDocQA_energy":97.26,"ViDoRe_syntheticDocQA_government_reports":98.28,"ViDoRe_syntheticDocQA_healthcare_industry":99.13,"ViDoRe_esg_reports_human_labeled_v2":87.5,"ViDoRe_biomedical_lectures_v2_multilingual":64.85,"ViDoRe_economics_reports_v2_multilingual":53.94,"ViDoRe_esg_reports_v2_multilingual":79.73,"VisRAG_ArxivQA":92.73,"VisRAG_ChartQA":95.07,"VisRAG_MP-DocVQA":87.31,"VisRAG_SlideVQA":95.9,"VisRAG_InfoVQA":93.0,"VisRAG_PlotQA":92.49,"ViDoSeek-doc":82.6,"MMLongBench-doc":52.03,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":50.29,"MMLongBench-page-old":28.18}
|
| 4 |
+
{"Rank":4,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-8B\">Qwen3-VL-Embedding-8B<\/a>","Model Size(B)":8.14,"Date":"2026-01-06","Visdoc-Overall":83.09,"ViDoRe-V1":87.21,"ViDoRe-V2":69.86,"VisRAG":88.68,"VisDoc-OOD":72.19,"ViDoRe_arxivqa":87.0,"ViDoRe_docvqa":54.02,"ViDoRe_infovqa":90.86,"ViDoRe_tabfquad":96.67,"ViDoRe_tatdqa":69.97,"ViDoRe_shiftproject":84.14,"ViDoRe_syntheticDocQA_artificial_intelligence":99.26,"ViDoRe_syntheticDocQA_energy":94.41,"ViDoRe_syntheticDocQA_government_reports":98.02,"ViDoRe_syntheticDocQA_healthcare_industry":97.79,"ViDoRe_esg_reports_human_labeled_v2":71.39,"ViDoRe_biomedical_lectures_v2_multilingual":71.59,"ViDoRe_economics_reports_v2_multilingual":67.18,"ViDoRe_esg_reports_v2_multilingual":69.28,"VisRAG_ArxivQA":88.21,"VisRAG_ChartQA":88.34,"VisRAG_MP-DocVQA":89.27,"VisRAG_SlideVQA":96.73,"VisRAG_InfoVQA":94.75,"VisRAG_PlotQA":74.8,"ViDoSeek-doc":86.1,"MMLongBench-doc":58.29,"ViDoSeek-page-fixed":88.38,"MMLongBench-page-fixed":60.29,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 5 |
+
{"Rank":5,"Models":"WeMM-Embedding-8B","Model Size(B)":8.77,"Date":"2025-12-16","Visdoc-Overall":82.49,"ViDoRe-V1":89.47,"ViDoRe-V2":59.31,"VisRAG":90.43,"VisDoc-OOD":70.13,"ViDoRe_arxivqa":92.22,"ViDoRe_docvqa":60.95,"ViDoRe_infovqa":92.83,"ViDoRe_tabfquad":93.41,"ViDoRe_tatdqa":77.03,"ViDoRe_shiftproject":88.5,"ViDoRe_syntheticDocQA_artificial_intelligence":98.15,"ViDoRe_syntheticDocQA_energy":96.21,"ViDoRe_syntheticDocQA_government_reports":96.17,"ViDoRe_syntheticDocQA_healthcare_industry":99.26,"ViDoRe_esg_reports_human_labeled_v2":62.97,"ViDoRe_biomedical_lectures_v2_multilingual":61.9,"ViDoRe_economics_reports_v2_multilingual":56.85,"ViDoRe_esg_reports_v2_multilingual":55.51,"VisRAG_ArxivQA":90.42,"VisRAG_ChartQA":91.4,"VisRAG_MP-DocVQA":91.82,"VisRAG_SlideVQA":97.0,"VisRAG_InfoVQA":95.97,"VisRAG_PlotQA":75.95,"ViDoSeek-doc":84.31,"MMLongBench-doc":55.95,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":23.34,"MMLongBench-page-old":17.05}
|
| 6 |
+
{"Rank":6,"Models":"<a href=\"https:\/\/github.com\/360CVGroup\/RzenEmbed\">RzenEmbed-v2-7B<\/a>","Model Size(B)":8.29,"Date":"2025-09-20","Visdoc-Overall":82.31,"ViDoRe-V1":89.7,"ViDoRe-V2":60.7,"VisRAG":88.7,"VisDoc-OOD":69.4,"ViDoRe_arxivqa":88.42,"ViDoRe_docvqa":59.61,"ViDoRe_infovqa":92.85,"ViDoRe_tabfquad":95.85,"ViDoRe_tatdqa":76.5,"ViDoRe_shiftproject":91.76,"ViDoRe_syntheticDocQA_artificial_intelligence":99.63,"ViDoRe_syntheticDocQA_energy":95.71,"ViDoRe_syntheticDocQA_government_reports":97.02,"ViDoRe_syntheticDocQA_healthcare_industry":99.63,"ViDoRe_esg_reports_human_labeled_v2":65.61,"ViDoRe_biomedical_lectures_v2_multilingual":61.71,"ViDoRe_economics_reports_v2_multilingual":59.1,"ViDoRe_esg_reports_v2_multilingual":56.39,"VisRAG_ArxivQA":89.38,"VisRAG_ChartQA":89.61,"VisRAG_MP-DocVQA":90.59,"VisRAG_SlideVQA":95.68,"VisRAG_InfoVQA":94.11,"VisRAG_PlotQA":72.81,"ViDoSeek-doc":84.02,"MMLongBench-doc":54.78,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":23.09,"MMLongBench-page-old":15.62}
|
| 7 |
+
{"Rank":7,"Models":"RzenEmbed-v1-7B","Model Size(B)":8.29,"Date":"2025-07-30","Visdoc-Overall":82.0,"ViDoRe-V1":89.47,"ViDoRe-V2":60.77,"VisRAG":87.92,"VisDoc-OOD":69.3,"ViDoRe_arxivqa":86.89,"ViDoRe_docvqa":57.99,"ViDoRe_infovqa":92.24,"ViDoRe_tabfquad":96.25,"ViDoRe_tatdqa":75.22,"ViDoRe_shiftproject":93.35,"ViDoRe_syntheticDocQA_artificial_intelligence":99.63,"ViDoRe_syntheticDocQA_energy":95.25,"ViDoRe_syntheticDocQA_government_reports":98.28,"ViDoRe_syntheticDocQA_healthcare_industry":99.63,"ViDoRe_esg_reports_human_labeled_v2":63.67,"ViDoRe_biomedical_lectures_v2_multilingual":61.58,"ViDoRe_economics_reports_v2_multilingual":60.8,"ViDoRe_esg_reports_v2_multilingual":57.05,"VisRAG_ArxivQA":86.97,"VisRAG_ChartQA":88.34,"VisRAG_MP-DocVQA":90.7,"VisRAG_SlideVQA":95.75,"VisRAG_InfoVQA":94.37,"VisRAG_PlotQA":71.39,"ViDoSeek-doc":83.92,"MMLongBench-doc":54.67,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":23.06,"MMLongBench-page-old":16.1}
|
| 8 |
+
{"Rank":8,"Models":"Taichu-UniRetriever-v1-2B","Model Size(B)":2.13,"Date":"2026-01-06","Visdoc-Overall":81.93,"ViDoRe-V1":89.81,"ViDoRe-V2":57.34,"VisRAG":88.79,"VisDoc-OOD":71.13,"ViDoRe_arxivqa":88.14,"ViDoRe_docvqa":59.09,"ViDoRe_infovqa":92.63,"ViDoRe_tabfquad":96.67,"ViDoRe_tatdqa":80.07,"ViDoRe_shiftproject":89.93,"ViDoRe_syntheticDocQA_artificial_intelligence":99.26,"ViDoRe_syntheticDocQA_energy":98.06,"ViDoRe_syntheticDocQA_government_reports":95.93,"ViDoRe_syntheticDocQA_healthcare_industry":98.32,"ViDoRe_esg_reports_human_labeled_v2":63.6,"ViDoRe_biomedical_lectures_v2_multilingual":62.47,"ViDoRe_economics_reports_v2_multilingual":51.21,"ViDoRe_esg_reports_v2_multilingual":52.08,"VisRAG_ArxivQA":87.25,"VisRAG_ChartQA":87.33,"VisRAG_MP-DocVQA":91.99,"VisRAG_SlideVQA":96.93,"VisRAG_InfoVQA":95.9,"VisRAG_PlotQA":73.32,"ViDoSeek-doc":85.07,"MMLongBench-doc":57.19,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":20.92,"MMLongBench-page-old":16.49}
|
| 9 |
+
{"Rank":9,"Models":"<a href=\"https:\/\/huggingface.co\/Haon-Chen\/e5-omni-7B\">e5-omni-7B<\/a>","Model Size(B)":8.0,"Date":"2026-01-09","Visdoc-Overall":81.3,"ViDoRe-V1":87.58,"ViDoRe-V2":62.39,"VisRAG":87.46,"VisDoc-OOD":69.21,"ViDoRe_arxivqa":87.57,"ViDoRe_docvqa":57.96,"ViDoRe_infovqa":92.29,"ViDoRe_tabfquad":93.19,"ViDoRe_tatdqa":70.9,"ViDoRe_shiftproject":85.79,"ViDoRe_syntheticDocQA_artificial_intelligence":98.15,"ViDoRe_syntheticDocQA_energy":94.95,"ViDoRe_syntheticDocQA_government_reports":96.28,"ViDoRe_syntheticDocQA_healthcare_industry":98.69,"ViDoRe_esg_reports_human_labeled_v2":64.03,"ViDoRe_biomedical_lectures_v2_multilingual":64.48,"ViDoRe_economics_reports_v2_multilingual":60.34,"ViDoRe_esg_reports_v2_multilingual":60.71,"VisRAG_ArxivQA":88.02,"VisRAG_ChartQA":89.02,"VisRAG_MP-DocVQA":89.07,"VisRAG_SlideVQA":95.99,"VisRAG_InfoVQA":93.54,"VisRAG_PlotQA":69.15,"ViDoSeek-doc":83.44,"MMLongBench-doc":54.98,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":23.11,"MMLongBench-page-old":15.78}
|
| 10 |
+
{"Rank":10,"Models":"<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen3-VL-Embedding-2B\">Qwen3-VL-Embedding-2B<\/a>","Model Size(B)":2.13,"Date":"2026-01-06","Visdoc-Overall":80.16,"ViDoRe-V1":84.37,"ViDoRe-V2":65.34,"VisRAG":86.43,"VisDoc-OOD":70.0,"ViDoRe_arxivqa":81.8,"ViDoRe_docvqa":49.36,"ViDoRe_infovqa":88.35,"ViDoRe_tabfquad":96.4,"ViDoRe_tatdqa":63.19,"ViDoRe_shiftproject":83.9,"ViDoRe_syntheticDocQA_artificial_intelligence":97.05,"ViDoRe_syntheticDocQA_energy":90.84,"ViDoRe_syntheticDocQA_government_reports":95.68,"ViDoRe_syntheticDocQA_healthcare_industry":97.09,"ViDoRe_esg_reports_human_labeled_v2":64.56,"ViDoRe_biomedical_lectures_v2_multilingual":68.23,"ViDoRe_economics_reports_v2_multilingual":63.09,"ViDoRe_esg_reports_v2_multilingual":65.49,"VisRAG_ArxivQA":83.58,"VisRAG_ChartQA":87.13,"VisRAG_MP-DocVQA":87.06,"VisRAG_SlideVQA":95.59,"VisRAG_InfoVQA":92.88,"VisRAG_PlotQA":72.33,"ViDoSeek-doc":85.0,"MMLongBench-doc":54.99,"ViDoSeek-page-fixed":81.94,"MMLongBench-page-fixed":55.73,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 11 |
+
{"Rank":11,"Models":"WeMM-Embedding-2B","Model Size(B)":2.13,"Date":"2025-12-16","Visdoc-Overall":79.85,"ViDoRe-V1":87.22,"ViDoRe-V2":53.4,"VisRAG":88.84,"VisDoc-OOD":68.89,"ViDoRe_arxivqa":90.25,"ViDoRe_docvqa":59.11,"ViDoRe_infovqa":91.2,"ViDoRe_tabfquad":89.41,"ViDoRe_tatdqa":73.81,"ViDoRe_shiftproject":82.61,"ViDoRe_syntheticDocQA_artificial_intelligence":98.15,"ViDoRe_syntheticDocQA_energy":94.08,"ViDoRe_syntheticDocQA_government_reports":95.35,"ViDoRe_syntheticDocQA_healthcare_industry":98.28,"ViDoRe_esg_reports_human_labeled_v2":54.54,"ViDoRe_biomedical_lectures_v2_multilingual":55.64,"ViDoRe_economics_reports_v2_multilingual":53.93,"ViDoRe_esg_reports_v2_multilingual":49.5,"VisRAG_ArxivQA":88.62,"VisRAG_ChartQA":88.61,"VisRAG_MP-DocVQA":91.01,"VisRAG_SlideVQA":95.83,"VisRAG_InfoVQA":93.74,"VisRAG_PlotQA":75.23,"ViDoSeek-doc":83.14,"MMLongBench-doc":54.65,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.27,"MMLongBench-page-old":16.53}
|
| 12 |
+
{"Rank":12,"Models":"RzenEmbed-v1-2B","Model Size(B)":2.21,"Date":"2025-07-16","Visdoc-Overall":79.4,"ViDoRe-V1":86.98,"ViDoRe-V2":57.63,"VisRAG":85.35,"VisDoc-OOD":67.11,"ViDoRe_arxivqa":83.4,"ViDoRe_docvqa":56.09,"ViDoRe_infovqa":89.89,"ViDoRe_tabfquad":94.39,"ViDoRe_tatdqa":69.32,"ViDoRe_shiftproject":89.88,"ViDoRe_syntheticDocQA_artificial_intelligence":98.76,"ViDoRe_syntheticDocQA_energy":92.78,"ViDoRe_syntheticDocQA_government_reports":97.19,"ViDoRe_syntheticDocQA_healthcare_industry":98.15,"ViDoRe_esg_reports_human_labeled_v2":58.93,"ViDoRe_biomedical_lectures_v2_multilingual":59.87,"ViDoRe_economics_reports_v2_multilingual":57.98,"ViDoRe_esg_reports_v2_multilingual":53.73,"VisRAG_ArxivQA":83.68,"VisRAG_ChartQA":86.39,"VisRAG_MP-DocVQA":85.54,"VisRAG_SlideVQA":94.07,"VisRAG_InfoVQA":91.92,"VisRAG_PlotQA":70.52,"ViDoSeek-doc":82.27,"MMLongBench-doc":51.95,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.89,"MMLongBench-page-old":16.21}
|
| 13 |
+
{"Rank":13,"Models":"<a href=\"https:\/\/seed1-6-embedding.github.io\">seed-1.6-embedding<\/a>","Model Size(B)":"unknown","Date":"2025-06-17","Visdoc-Overall":78.37,"ViDoRe-V1":85.53,"ViDoRe-V2":56.57,"VisRAG":84.74,"VisDoc-OOD":67.09,"ViDoRe_arxivqa":88.46,"ViDoRe_docvqa":60.36,"ViDoRe_infovqa":90.62,"ViDoRe_tabfquad":87.82,"ViDoRe_tatdqa":74.98,"ViDoRe_shiftproject":73.34,"ViDoRe_syntheticDocQA_artificial_intelligence":96.32,"ViDoRe_syntheticDocQA_energy":92.83,"ViDoRe_syntheticDocQA_government_reports":93.54,"ViDoRe_syntheticDocQA_healthcare_industry":97.05,"ViDoRe_esg_reports_human_labeled_v2":63.3,"ViDoRe_biomedical_lectures_v2_multilingual":57.14,"ViDoRe_economics_reports_v2_multilingual":53.85,"ViDoRe_esg_reports_v2_multilingual":51.99,"VisRAG_ArxivQA":85.79,"VisRAG_ChartQA":81.88,"VisRAG_MP-DocVQA":89.43,"VisRAG_SlideVQA":94.9,"VisRAG_InfoVQA":92.41,"VisRAG_PlotQA":64.02,"ViDoSeek-doc":82.57,"MMLongBench-doc":51.61,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.8,"MMLongBench-page-old":15.57}
|
| 14 |
+
{"Rank":14,"Models":"<a href=\"https:\/\/huggingface.co\/Haon-Chen\/e5-omni-3B\">e5-omni-3B<\/a>","Model Size(B)":3.0,"Date":"2026-01-09","Visdoc-Overall":78.11,"ViDoRe-V1":83.13,"ViDoRe-V2":59.53,"VisRAG":85.48,"VisDoc-OOD":68.07,"ViDoRe_arxivqa":82.2,"ViDoRe_docvqa":51.99,"ViDoRe_infovqa":88.88,"ViDoRe_tabfquad":88.86,"ViDoRe_tatdqa":58.78,"ViDoRe_shiftproject":80.14,"ViDoRe_syntheticDocQA_artificial_intelligence":98.15,"ViDoRe_syntheticDocQA_energy":93.17,"ViDoRe_syntheticDocQA_government_reports":93.21,"ViDoRe_syntheticDocQA_healthcare_industry":95.92,"ViDoRe_esg_reports_human_labeled_v2":61.84,"ViDoRe_biomedical_lectures_v2_multilingual":60.92,"ViDoRe_economics_reports_v2_multilingual":64.39,"ViDoRe_esg_reports_v2_multilingual":50.97,"VisRAG_ArxivQA":84.52,"VisRAG_ChartQA":86.75,"VisRAG_MP-DocVQA":85.51,"VisRAG_SlideVQA":95.4,"VisRAG_InfoVQA":92.55,"VisRAG_PlotQA":68.17,"ViDoSeek-doc":83.29,"MMLongBench-doc":52.85,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.38,"MMLongBench-page-old":15.0}
|
| 15 |
+
{"Rank":15,"Models":"<a href=\"https:\/\/huggingface.co\/OpenSearch-AI\/Ops-MM-embedding-v1-7B\">Ops-MM-embedding-v1-7B<\/a>","Model Size(B)":8.29,"Date":"2025-07-02","Visdoc-Overall":74.99,"ViDoRe-V1":80.05,"ViDoRe-V2":59.59,"VisRAG":79.31,"VisDoc-OOD":67.49,"ViDoRe_arxivqa":78.17,"ViDoRe_docvqa":49.11,"ViDoRe_infovqa":86.59,"ViDoRe_tabfquad":91.4,"ViDoRe_tatdqa":55.59,"ViDoRe_shiftproject":76.17,"ViDoRe_syntheticDocQA_artificial_intelligence":90.99,"ViDoRe_syntheticDocQA_energy":87.53,"ViDoRe_syntheticDocQA_government_reports":91.56,"ViDoRe_syntheticDocQA_healthcare_industry":93.39,"ViDoRe_esg_reports_human_labeled_v2":66.27,"ViDoRe_biomedical_lectures_v2_multilingual":54.34,"ViDoRe_economics_reports_v2_multilingual":60.92,"ViDoRe_esg_reports_v2_multilingual":56.82,"VisRAG_ArxivQA":78.18,"VisRAG_ChartQA":79.53,"VisRAG_MP-DocVQA":78.2,"VisRAG_SlideVQA":91.11,"VisRAG_InfoVQA":87.16,"VisRAG_PlotQA":61.71,"ViDoSeek-doc":83.47,"MMLongBench-doc":51.51,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.47,"MMLongBench-page-old":15.89}
|
| 16 |
+
{"Rank":16,"Models":"<a href=\"https:\/\/huggingface.co\/Alibaba-NLP\/gme-Qwen2-VL-7B-Instruct\">gme-Qwen2-VL-7B-Instruct<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Visdoc-Overall":73.96,"ViDoRe-V1":89.55,"ViDoRe-V2":55.47,"VisRAG":84.95,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":87.53,"ViDoRe_docvqa":56.59,"ViDoRe_infovqa":92.17,"ViDoRe_tabfquad":92.71,"ViDoRe_tatdqa":76.58,"ViDoRe_shiftproject":95.64,"ViDoRe_syntheticDocQA_artificial_intelligence":99.63,"ViDoRe_syntheticDocQA_energy":95.69,"ViDoRe_syntheticDocQA_government_reports":99.5,"ViDoRe_syntheticDocQA_healthcare_industry":99.5,"ViDoRe_esg_reports_human_labeled_v2":62.76,"ViDoRe_biomedical_lectures_v2_multilingual":49.78,"ViDoRe_economics_reports_v2_multilingual":53.91,"ViDoRe_esg_reports_v2_multilingual":55.43,"VisRAG_ArxivQA":87.69,"VisRAG_ChartQA":81.29,"VisRAG_MP-DocVQA":89.08,"VisRAG_SlideVQA":94.72,"VisRAG_InfoVQA":93.49,"VisRAG_PlotQA":63.44,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":83.57,"MMLongBench-page-fixed":55.32,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 17 |
+
{"Rank":17,"Models":"Crotchet-embedding-2B","Model Size(B)":2.13,"Date":"2025-12-31","Visdoc-Overall":73.59,"ViDoRe-V1":78.16,"ViDoRe-V2":53.28,"VisRAG":82.21,"VisDoc-OOD":65.51,"ViDoRe_arxivqa":84.34,"ViDoRe_docvqa":42.93,"ViDoRe_infovqa":85.18,"ViDoRe_tabfquad":93.29,"ViDoRe_tatdqa":53.04,"ViDoRe_shiftproject":71.02,"ViDoRe_syntheticDocQA_artificial_intelligence":90.74,"ViDoRe_syntheticDocQA_energy":85.07,"ViDoRe_syntheticDocQA_government_reports":85.45,"ViDoRe_syntheticDocQA_healthcare_industry":90.59,"ViDoRe_esg_reports_human_labeled_v2":62.65,"ViDoRe_biomedical_lectures_v2_multilingual":46.92,"ViDoRe_economics_reports_v2_multilingual":51.47,"ViDoRe_esg_reports_v2_multilingual":52.09,"VisRAG_ArxivQA":81.64,"VisRAG_ChartQA":85.47,"VisRAG_MP-DocVQA":76.0,"VisRAG_SlideVQA":90.85,"VisRAG_InfoVQA":88.54,"VisRAG_PlotQA":70.74,"ViDoSeek-doc":81.84,"MMLongBench-doc":49.18,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":78.36,"MMLongBench-page-old":51.39}
|
| 18 |
+
{"Rank":18,"Models":"<a href=\"https:\/\/huggingface.co\/Alibaba-NLP\/gme-Qwen2-VL-2B-Instruct\">gme-Qwen2-VL-2B-Instruct<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Visdoc-Overall":71.88,"ViDoRe-V1":87.12,"ViDoRe-V2":53.88,"VisRAG":82.43,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":82.53,"ViDoRe_docvqa":55.16,"ViDoRe_infovqa":90.69,"ViDoRe_tabfquad":93.31,"ViDoRe_tatdqa":70.26,"ViDoRe_shiftproject":92.85,"ViDoRe_syntheticDocQA_artificial_intelligence":98.13,"ViDoRe_syntheticDocQA_energy":92.56,"ViDoRe_syntheticDocQA_government_reports":97.15,"ViDoRe_syntheticDocQA_healthcare_industry":98.52,"ViDoRe_esg_reports_human_labeled_v2":60.29,"ViDoRe_biomedical_lectures_v2_multilingual":53.99,"ViDoRe_economics_reports_v2_multilingual":50.8,"ViDoRe_esg_reports_v2_multilingual":50.43,"VisRAG_ArxivQA":81.95,"VisRAG_ChartQA":79.09,"VisRAG_MP-DocVQA":84.34,"VisRAG_SlideVQA":93.68,"VisRAG_InfoVQA":91.41,"VisRAG_PlotQA":64.12,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":79.81,"MMLongBench-page-fixed":51.32,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 19 |
+
{"Rank":19,"Models":"<a href=\"https:\/\/huggingface.co\/zhibinlan\/UME-R1-7B\">UME-R1-7B<\/a>","Model Size(B)":8.29,"Date":"2025-10-20","Visdoc-Overall":71.71,"ViDoRe-V1":75.66,"ViDoRe-V2":50.53,"VisRAG":83.7,"VisDoc-OOD":58.3,"ViDoRe_arxivqa":73.56,"ViDoRe_docvqa":41.1,"ViDoRe_infovqa":80.83,"ViDoRe_tabfquad":90.24,"ViDoRe_tatdqa":46.67,"ViDoRe_shiftproject":64.96,"ViDoRe_syntheticDocQA_artificial_intelligence":89.47,"ViDoRe_syntheticDocQA_energy":85.7,"ViDoRe_syntheticDocQA_government_reports":89.82,"ViDoRe_syntheticDocQA_healthcare_industry":94.29,"ViDoRe_esg_reports_human_labeled_v2":50.41,"ViDoRe_biomedical_lectures_v2_multilingual":50.73,"ViDoRe_economics_reports_v2_multilingual":57.8,"ViDoRe_esg_reports_v2_multilingual":43.18,"VisRAG_ArxivQA":80.45,"VisRAG_ChartQA":84.95,"VisRAG_MP-DocVQA":83.43,"VisRAG_SlideVQA":91.47,"VisRAG_InfoVQA":89.2,"VisRAG_PlotQA":72.7,"ViDoSeek-doc":75.33,"MMLongBench-doc":41.28,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":21.28,"MMLongBench-page-old":12.32}
|
| 20 |
+
{"Rank":20,"Models":"<a href=\"https:\/\/huggingface.co\/OpenSearch-AI\/Ops-MM-embedding-v1-2B\">Ops-MM-embedding-v1-2B<\/a>","Model Size(B)":2.21,"Date":"2025-07-02","Visdoc-Overall":71.48,"ViDoRe-V1":76.39,"ViDoRe-V2":53.18,"VisRAG":77.64,"VisDoc-OOD":65.09,"ViDoRe_arxivqa":73.72,"ViDoRe_docvqa":45.01,"ViDoRe_infovqa":81.41,"ViDoRe_tabfquad":88.81,"ViDoRe_tatdqa":49.59,"ViDoRe_shiftproject":72.57,"ViDoRe_syntheticDocQA_artificial_intelligence":89.8,"ViDoRe_syntheticDocQA_energy":84.27,"ViDoRe_syntheticDocQA_government_reports":87.06,"ViDoRe_syntheticDocQA_healthcare_industry":91.64,"ViDoRe_esg_reports_human_labeled_v2":58.57,"ViDoRe_biomedical_lectures_v2_multilingual":52.87,"ViDoRe_economics_reports_v2_multilingual":47.89,"ViDoRe_esg_reports_v2_multilingual":53.39,"VisRAG_ArxivQA":73.84,"VisRAG_ChartQA":80.17,"VisRAG_MP-DocVQA":75.06,"VisRAG_SlideVQA":89.15,"VisRAG_InfoVQA":86.66,"VisRAG_PlotQA":60.99,"ViDoSeek-doc":82.29,"MMLongBench-doc":47.89,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":21.44,"MMLongBench-page-old":13.06}
|
| 21 |
+
{"Rank":21,"Models":"colpali-v1.3","Model Size(B)":"unknown","Date":"2025-12-29","Visdoc-Overall":70.5,"ViDoRe-V1":84.57,"ViDoRe-V2":54.81,"VisRAG":81.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":83.33,"ViDoRe_docvqa":59.2,"ViDoRe_infovqa":85.38,"ViDoRe_tabfquad":86.63,"ViDoRe_tatdqa":70.47,"ViDoRe_shiftproject":76.83,"ViDoRe_syntheticDocQA_artificial_intelligence":96.26,"ViDoRe_syntheticDocQA_energy":95.02,"ViDoRe_syntheticDocQA_government_reports":95.59,"ViDoRe_syntheticDocQA_healthcare_industry":97.02,"ViDoRe_esg_reports_human_labeled_v2":57.25,"ViDoRe_biomedical_lectures_v2_multilingual":56.25,"ViDoRe_economics_reports_v2_multilingual":50.2,"ViDoRe_esg_reports_v2_multilingual":55.56,"VisRAG_ArxivQA":80.9,"VisRAG_ChartQA":77.04,"VisRAG_MP-DocVQA":87.03,"VisRAG_SlideVQA":95.13,"VisRAG_InfoVQA":85.73,"VisRAG_PlotQA":60.18,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":79.22,"MMLongBench-page-fixed":53.18,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 22 |
+
{"Rank":22,"Models":"<a href=\"https:\/\/huggingface.co\/zhibinlan\/UME-R1-2B\">UME-R1-2B<\/a>","Model Size(B)":2.21,"Date":"2025-10-20","Visdoc-Overall":68.16,"ViDoRe-V1":72.39,"ViDoRe-V2":46.16,"VisRAG":79.22,"VisDoc-OOD":57.78,"ViDoRe_arxivqa":73.86,"ViDoRe_docvqa":37.85,"ViDoRe_infovqa":76.19,"ViDoRe_tabfquad":86.11,"ViDoRe_tatdqa":40.63,"ViDoRe_shiftproject":66.79,"ViDoRe_syntheticDocQA_artificial_intelligence":85.87,"ViDoRe_syntheticDocQA_energy":83.25,"ViDoRe_syntheticDocQA_government_reports":82.56,"ViDoRe_syntheticDocQA_healthcare_industry":90.83,"ViDoRe_esg_reports_human_labeled_v2":50.16,"ViDoRe_biomedical_lectures_v2_multilingual":46.15,"ViDoRe_economics_reports_v2_multilingual":45.69,"ViDoRe_esg_reports_v2_multilingual":42.65,"VisRAG_ArxivQA":74.26,"VisRAG_ChartQA":85.98,"VisRAG_MP-DocVQA":75.59,"VisRAG_SlideVQA":87.11,"VisRAG_InfoVQA":84.38,"VisRAG_PlotQA":67.98,"ViDoSeek-doc":75.89,"MMLongBench-doc":39.68,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":21.2,"MMLongBench-page-old":11.9}
|
| 23 |
+
{"Rank":23,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-7B<\/a>","Model Size(B)":8.03,"Date":"2025-07-06","Visdoc-Overall":68.1,"ViDoRe-V1":70.68,"ViDoRe-V2":49.57,"VisRAG":79.45,"VisDoc-OOD":58.21,"ViDoRe_arxivqa":73.29,"ViDoRe_docvqa":38.27,"ViDoRe_infovqa":80.59,"ViDoRe_tabfquad":80.7,"ViDoRe_tatdqa":37.79,"ViDoRe_shiftproject":52.02,"ViDoRe_syntheticDocQA_artificial_intelligence":85.99,"ViDoRe_syntheticDocQA_energy":84.78,"ViDoRe_syntheticDocQA_government_reports":84.96,"ViDoRe_syntheticDocQA_healthcare_industry":88.37,"ViDoRe_esg_reports_human_labeled_v2":50.67,"ViDoRe_biomedical_lectures_v2_multilingual":50.89,"ViDoRe_economics_reports_v2_multilingual":54.38,"ViDoRe_esg_reports_v2_multilingual":42.33,"VisRAG_ArxivQA":73.96,"VisRAG_ChartQA":82.71,"VisRAG_MP-DocVQA":75.15,"VisRAG_SlideVQA":87.58,"VisRAG_InfoVQA":87.91,"VisRAG_PlotQA":69.42,"ViDoSeek-doc":73.82,"MMLongBench-doc":42.61,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.52,"MMLongBench-page-old":13.32}
|
| 24 |
+
{"Rank":24,"Models":"<a href=\"https:\/\/huggingface.co\/VLM2Vec\/VLM2Vec-V2.0\">VLM2Vec-V2.0-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Visdoc-Overall":63.56,"ViDoRe-V1":74.39,"ViDoRe-V2":44.61,"VisRAG":79.33,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":78.87,"ViDoRe_docvqa":37.06,"ViDoRe_infovqa":82.71,"ViDoRe_tabfquad":87.77,"ViDoRe_tatdqa":44.29,"ViDoRe_shiftproject":61.04,"ViDoRe_syntheticDocQA_artificial_intelligence":89.12,"ViDoRe_syntheticDocQA_energy":86.29,"ViDoRe_syntheticDocQA_government_reports":85.58,"ViDoRe_syntheticDocQA_healthcare_industry":91.13,"ViDoRe_esg_reports_human_labeled_v2":45.79,"ViDoRe_biomedical_lectures_v2_multilingual":44.59,"ViDoRe_economics_reports_v2_multilingual":42.33,"ViDoRe_esg_reports_v2_multilingual":45.74,"VisRAG_ArxivQA":76.74,"VisRAG_ChartQA":84.2,"VisRAG_MP-DocVQA":71.8,"VisRAG_SlideVQA":91.44,"VisRAG_InfoVQA":85.88,"VisRAG_PlotQA":65.91,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":80.3,"MMLongBench-page-fixed":44.7,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 25 |
+
{"Rank":25,"Models":"<a href=\"https:\/\/github.com\/GaryGuTC\/UniME-v2\">UniME-V2-LLaVA-OneVision-7B<\/a>","Model Size(B)":8.03,"Date":"2025-10-15","Visdoc-Overall":60.24,"ViDoRe-V1":61.76,"ViDoRe-V2":42.0,"VisRAG":70.53,"VisDoc-OOD":58.3,"ViDoRe_arxivqa":51.9,"ViDoRe_docvqa":38.17,"ViDoRe_infovqa":73.16,"ViDoRe_tabfquad":57.75,"ViDoRe_tatdqa":35.48,"ViDoRe_shiftproject":45.42,"ViDoRe_syntheticDocQA_artificial_intelligence":76.85,"ViDoRe_syntheticDocQA_energy":77.27,"ViDoRe_syntheticDocQA_government_reports":79.87,"ViDoRe_syntheticDocQA_healthcare_industry":81.71,"ViDoRe_esg_reports_human_labeled_v2":54.67,"ViDoRe_biomedical_lectures_v2_multilingual":33.75,"ViDoRe_economics_reports_v2_multilingual":35.95,"ViDoRe_esg_reports_v2_multilingual":43.63,"VisRAG_ArxivQA":53.13,"VisRAG_ChartQA":83.67,"VisRAG_MP-DocVQA":66.4,"VisRAG_SlideVQA":86.44,"VisRAG_InfoVQA":82.58,"VisRAG_PlotQA":50.96,"ViDoSeek-doc":75.79,"MMLongBench-doc":40.8,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":22.89,"MMLongBench-page-old":11.96}
|
| 26 |
+
{"Rank":26,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2503.19900\">interestFM-UIR-CAFe-0.5B<\/a>","Model Size(B)":0.894,"Date":"2025-07-06","Visdoc-Overall":54.83,"ViDoRe-V1":56.93,"ViDoRe-V2":32.59,"VisRAG":68.57,"VisDoc-OOD":47.59,"ViDoRe_arxivqa":51.19,"ViDoRe_docvqa":25.46,"ViDoRe_infovqa":71.95,"ViDoRe_tabfquad":59.64,"ViDoRe_tatdqa":27.05,"ViDoRe_shiftproject":31.8,"ViDoRe_syntheticDocQA_artificial_intelligence":78.74,"ViDoRe_syntheticDocQA_energy":70.75,"ViDoRe_syntheticDocQA_government_reports":75.47,"ViDoRe_syntheticDocQA_healthcare_industry":77.29,"ViDoRe_esg_reports_human_labeled_v2":36.24,"ViDoRe_biomedical_lectures_v2_multilingual":29.72,"ViDoRe_economics_reports_v2_multilingual":37.61,"ViDoRe_esg_reports_v2_multilingual":26.78,"VisRAG_ArxivQA":57.82,"VisRAG_ChartQA":75.62,"VisRAG_MP-DocVQA":63.24,"VisRAG_SlideVQA":82.32,"VisRAG_InfoVQA":80.28,"VisRAG_PlotQA":52.14,"ViDoSeek-doc":61.41,"MMLongBench-doc":33.78,"ViDoSeek-page-fixed":"\u26a0\ufe0f! Please fix this score!","MMLongBench-page-fixed":"\u26a0\ufe0f! Please fix this score!","ViDoSeek-page-old":17.62,"MMLongBench-page-old":9.95}
|
| 27 |
+
{"Rank":27,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret-Qwen2.5VL-7b\">LamRA-Ret-Qwen2.5VL-7b<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Visdoc-Overall":47.61,"ViDoRe-V1":56.4,"ViDoRe-V2":33.62,"VisRAG":58.15,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":53.04,"ViDoRe_docvqa":25.42,"ViDoRe_infovqa":72.19,"ViDoRe_tabfquad":66.51,"ViDoRe_tatdqa":26.01,"ViDoRe_shiftproject":28.13,"ViDoRe_syntheticDocQA_artificial_intelligence":70.31,"ViDoRe_syntheticDocQA_energy":66.33,"ViDoRe_syntheticDocQA_government_reports":72.21,"ViDoRe_syntheticDocQA_healthcare_industry":83.8,"ViDoRe_esg_reports_human_labeled_v2":32.98,"ViDoRe_biomedical_lectures_v2_multilingual":35.87,"ViDoRe_economics_reports_v2_multilingual":32.37,"ViDoRe_esg_reports_v2_multilingual":33.26,"VisRAG_ArxivQA":38.03,"VisRAG_ChartQA":65.21,"VisRAG_MP-DocVQA":54.14,"VisRAG_SlideVQA":76.55,"VisRAG_InfoVQA":73.5,"VisRAG_PlotQA":41.46,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":68.54,"MMLongBench-page-fixed":35.83,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 28 |
+
{"Rank":28,"Models":"<a href=\"https:\/\/huggingface.co\/code-kunkun\/LamRA-Ret\">LamRA-Ret<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Visdoc-Overall":27.72,"ViDoRe-V1":33.83,"ViDoRe-V2":11.53,"VisRAG":37.56,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":31.54,"ViDoRe_docvqa":19.85,"ViDoRe_infovqa":63.67,"ViDoRe_tabfquad":53.47,"ViDoRe_tatdqa":7.88,"ViDoRe_shiftproject":15.95,"ViDoRe_syntheticDocQA_artificial_intelligence":29.83,"ViDoRe_syntheticDocQA_energy":36.05,"ViDoRe_syntheticDocQA_government_reports":41.18,"ViDoRe_syntheticDocQA_healthcare_industry":38.83,"ViDoRe_esg_reports_human_labeled_v2":6.91,"ViDoRe_biomedical_lectures_v2_multilingual":13.42,"ViDoRe_economics_reports_v2_multilingual":19.37,"ViDoRe_esg_reports_v2_multilingual":6.43,"VisRAG_ArxivQA":2.04,"VisRAG_ChartQA":42.74,"VisRAG_MP-DocVQA":33.37,"VisRAG_SlideVQA":56.3,"VisRAG_InfoVQA":56.86,"VisRAG_PlotQA":34.06,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":34.48,"MMLongBench-page-fixed":18.29,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 29 |
+
{"Rank":29,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-7B\">VLM2Vec-V1-Qwen2VL-7B<\/a>","Model Size(B)":8.29,"Date":"2025-12-29","Visdoc-Overall":26.82,"ViDoRe-V1":19.99,"ViDoRe-V2":9.24,"VisRAG":58.85,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":28.16,"ViDoRe_docvqa":18.98,"ViDoRe_infovqa":44.77,"ViDoRe_tabfquad":16.98,"ViDoRe_tatdqa":5.73,"ViDoRe_shiftproject":1.63,"ViDoRe_syntheticDocQA_artificial_intelligence":18.23,"ViDoRe_syntheticDocQA_energy":23.9,"ViDoRe_syntheticDocQA_government_reports":13.88,"ViDoRe_syntheticDocQA_healthcare_industry":27.61,"ViDoRe_esg_reports_human_labeled_v2":6.95,"ViDoRe_biomedical_lectures_v2_multilingual":5.22,"ViDoRe_economics_reports_v2_multilingual":13.75,"ViDoRe_esg_reports_v2_multilingual":11.05,"VisRAG_ArxivQA":52.85,"VisRAG_ChartQA":68.96,"VisRAG_MP-DocVQA":52.74,"VisRAG_SlideVQA":72.76,"VisRAG_InfoVQA":71.3,"VisRAG_PlotQA":34.52,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":77.4,"MMLongBench-page-fixed":36.76,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 30 |
+
{"Rank":30,"Models":"<a href=\"https:\/\/huggingface.co\/TIGER-Lab\/VLM2Vec-Qwen2VL-2B\">VLM2Vec-V1-Qwen2VL-2B<\/a>","Model Size(B)":2.21,"Date":"2025-12-29","Visdoc-Overall":26.0,"ViDoRe-V1":20.59,"ViDoRe-V2":13.19,"VisRAG":52.23,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":18.12,"ViDoRe_docvqa":14.01,"ViDoRe_infovqa":39.53,"ViDoRe_tabfquad":36.03,"ViDoRe_tatdqa":10.49,"ViDoRe_shiftproject":8.41,"ViDoRe_syntheticDocQA_artificial_intelligence":17.04,"ViDoRe_syntheticDocQA_energy":16.35,"ViDoRe_syntheticDocQA_government_reports":25.15,"ViDoRe_syntheticDocQA_healthcare_industry":20.79,"ViDoRe_esg_reports_human_labeled_v2":13.05,"ViDoRe_biomedical_lectures_v2_multilingual":6.51,"ViDoRe_economics_reports_v2_multilingual":12.86,"ViDoRe_esg_reports_v2_multilingual":20.32,"VisRAG_ArxivQA":41.22,"VisRAG_ChartQA":59.48,"VisRAG_MP-DocVQA":43.55,"VisRAG_SlideVQA":74.5,"VisRAG_InfoVQA":71.1,"VisRAG_PlotQA":23.51,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":67.81,"MMLongBench-page-fixed":26.04,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 31 |
+
{"Rank":31,"Models":"dp-embedding-v3-lite","Model Size(B)":8.29,"Date":"2025-12-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 32 |
+
{"Rank":32,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-2B<\/a>","Model Size(B)":2.21,"Date":"2025-11-24","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 33 |
+
{"Rank":33,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-3B<\/a>","Model Size(B)":3.75,"Date":"2025-11-24","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 34 |
+
{"Rank":34,"Models":"TCE-v1","Model Size(B)":8.0,"Date":"2025-10-31","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 35 |
+
{"Rank":35,"Models":"UniVec-CoT-7B","Model Size(B)":8.29,"Date":"unknown","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 36 |
+
{"Rank":36,"Models":"<a href=\"https:\/\/arxiv.org\/abs\/2511.19278\">ReMatch-7B<\/a>","Model Size(B)":8.29,"Date":"2025-11-24","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 37 |
+
{"Rank":37,"Models":"UniVec-7B","Model Size(B)":8.29,"Date":"unknown","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 38 |
+
{"Rank":38,"Models":"ReCo-7B","Model Size(B)":8.29,"Date":"2025-08-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 39 |
+
{"Rank":39,"Models":"OEmbedding-v1-7B","Model Size(B)":8.29,"Date":"2025-10-14","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 40 |
+
{"Rank":40,"Models":"<a href=\"https:\/\/github.com\/QQ-MM\/QQMM-embed\">QQMM-embed-v2<\/a>","Model Size(B)":8.29,"Date":"2025-09-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 41 |
+
{"Rank":41,"Models":"QQMM-embed-v3","Model Size(B)":8.29,"Date":"2025-12-30","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
| 42 |
+
{"Rank":42,"Models":"TCR","Model Size(B)":8.77,"Date":"2026-01-15","Visdoc-Overall":0.0,"ViDoRe-V1":0.0,"ViDoRe-V2":0.0,"VisRAG":0.0,"VisDoc-OOD":0.0,"ViDoRe_arxivqa":0.0,"ViDoRe_docvqa":0.0,"ViDoRe_infovqa":0.0,"ViDoRe_tabfquad":0.0,"ViDoRe_tatdqa":0.0,"ViDoRe_shiftproject":0.0,"ViDoRe_syntheticDocQA_artificial_intelligence":0.0,"ViDoRe_syntheticDocQA_energy":0.0,"ViDoRe_syntheticDocQA_government_reports":0.0,"ViDoRe_syntheticDocQA_healthcare_industry":0.0,"ViDoRe_esg_reports_human_labeled_v2":0.0,"ViDoRe_biomedical_lectures_v2_multilingual":0.0,"ViDoRe_economics_reports_v2_multilingual":0.0,"ViDoRe_esg_reports_v2_multilingual":0.0,"VisRAG_ArxivQA":0.0,"VisRAG_ChartQA":0.0,"VisRAG_MP-DocVQA":0.0,"VisRAG_SlideVQA":0.0,"VisRAG_InfoVQA":0.0,"VisRAG_PlotQA":0.0,"ViDoSeek-doc":0.0,"MMLongBench-doc":0.0,"ViDoSeek-page-fixed":0.0,"MMLongBench-page-fixed":0.0,"ViDoSeek-page-old":"\u2705","MMLongBench-page-old":"\u2705"}
|
utils_v2.py
CHANGED
|
@@ -6,29 +6,29 @@ from utils import create_hyperlinked_names, process_model_size
|
|
| 6 |
from datasets import *
|
| 7 |
|
| 8 |
BASE_COLS = ['Rank', 'Models', 'Model Size(B)', 'Date']
|
| 9 |
-
BASE_DATA_TITLE_TYPE = ['number', 'markdown', 'str', '
|
| 10 |
|
| 11 |
COLUMN_NAMES = BASE_COLS + ["Overall", 'Image-Overall', 'Video-Overall', 'Visdoc-Overall']
|
| 12 |
DATA_TITLE_TYPE = BASE_DATA_TITLE_TYPE + \
|
| 13 |
-
['number'] *
|
| 14 |
|
| 15 |
SUB_TASKS_I = ["I-CLS", "I-QA", "I-RET", "I-VG"]
|
| 16 |
TASKS_I = ['Image-Overall'] + SUB_TASKS_I + ALL_DATASETS_SPLITS['image']
|
| 17 |
COLUMN_NAMES_I = BASE_COLS + TASKS_I
|
| 18 |
DATA_TITLE_TYPE_I = BASE_DATA_TITLE_TYPE + \
|
| 19 |
-
['number'] * len(TASKS_I
|
| 20 |
|
| 21 |
SUB_TASKS_V = ["V-CLS", "V-QA", "V-RET", "V-MRET"]
|
| 22 |
TASKS_V = ['Video-Overall'] + SUB_TASKS_V + ALL_DATASETS_SPLITS['video']
|
| 23 |
COLUMN_NAMES_V = BASE_COLS + TASKS_V
|
| 24 |
DATA_TITLE_TYPE_V = BASE_DATA_TITLE_TYPE + \
|
| 25 |
-
['number'] * len(TASKS_V
|
| 26 |
|
| 27 |
SUB_TASKS_D = ['ViDoRe-V1', 'ViDoRe-V2', 'VisRAG', 'VisDoc-OOD']
|
| 28 |
TASKS_D = ['Visdoc-Overall'] + SUB_TASKS_D + ALL_DATASETS_SPLITS['visdoc']
|
| 29 |
COLUMN_NAMES_D = BASE_COLS + TASKS_D
|
| 30 |
DATA_TITLE_TYPE_D = BASE_DATA_TITLE_TYPE + \
|
| 31 |
-
['number'] * len(TASKS_D
|
| 32 |
|
| 33 |
TABLE_INTRODUCTION = """**MMEB**: Massive MultiModal Embedding Benchmark \n
|
| 34 |
Models are ranked based on **Overall**"""
|
|
@@ -41,7 +41,7 @@ TABLE_INTRODUCTION_V = """**V-CLS**: Video Classification, **V-QA**: (Video) Vis
|
|
| 41 |
Models are ranked based on **Video-Overall**"""
|
| 42 |
TABLE_INTRODUCTION_D = """**VisDoc**: Visual Document Understanding \n
|
| 43 |
Models are ranked based on **Visdoc-Overall** \n
|
| 44 |
-
**Attention Please! We have fixed the error found in ViDoSeek-page and MMLongBench-page (See this [GitHub issue](https://github.com/TIGER-AI-Lab/VLM2Vec/issues/167) for more info). We kindly ask the authors to rerun their models on the visdoc datasets and
|
| 45 |
|
| 46 |
LEADERBOARD_INFO = """
|
| 47 |
## Dataset Summary
|
|
@@ -76,6 +76,10 @@ def load_scores(raw_scores=None):
|
|
| 76 |
all_scores = {}
|
| 77 |
for modality, datasets_list in DATASETS.items(): # Ex.: ('image', {'I-CLS': [...], 'I-QA': [...]})
|
| 78 |
for sub_task, datasets in datasets_list.items(): # Ex.: ('I-CLS', ['VOC2007', 'N24News', ...])
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
for dataset in datasets: # Ex.: 'VOC2007'
|
| 80 |
score = raw_scores.get(modality, {}).get(dataset, 0.0)
|
| 81 |
score = 0.0 if isinstance(score, str) and "N/A" in score else score
|
|
@@ -173,12 +177,13 @@ def search_and_filter_models(df, query, min_size, max_size):
|
|
| 173 |
|
| 174 |
return filtered_df[COLUMN_NAMES]
|
| 175 |
|
| 176 |
-
def save_ranking_summary(df, name, dir='rankings'):
|
| 177 |
csv_path, json_path = os.path.join(dir, f'{name}.csv'), os.path.join(dir, f'{name}.jsonl')
|
| 178 |
-
|
| 179 |
-
|
|
|
|
| 180 |
return csv_path, json_path
|
| 181 |
|
| 182 |
def download_ranking(df, name, format='csv', dir='rankings'):
|
| 183 |
-
csv_path, json_path = save_ranking_summary(df, name, dir)
|
| 184 |
return csv_path if format == 'csv' else json_path
|
|
|
|
| 6 |
from datasets import *
|
| 7 |
|
| 8 |
BASE_COLS = ['Rank', 'Models', 'Model Size(B)', 'Date']
|
| 9 |
+
BASE_DATA_TITLE_TYPE = ['number', 'markdown', 'str', 'str']
|
| 10 |
|
| 11 |
COLUMN_NAMES = BASE_COLS + ["Overall", 'Image-Overall', 'Video-Overall', 'Visdoc-Overall']
|
| 12 |
DATA_TITLE_TYPE = BASE_DATA_TITLE_TYPE + \
|
| 13 |
+
['number'] * 4
|
| 14 |
|
| 15 |
SUB_TASKS_I = ["I-CLS", "I-QA", "I-RET", "I-VG"]
|
| 16 |
TASKS_I = ['Image-Overall'] + SUB_TASKS_I + ALL_DATASETS_SPLITS['image']
|
| 17 |
COLUMN_NAMES_I = BASE_COLS + TASKS_I
|
| 18 |
DATA_TITLE_TYPE_I = BASE_DATA_TITLE_TYPE + \
|
| 19 |
+
['number'] * len(TASKS_I)
|
| 20 |
|
| 21 |
SUB_TASKS_V = ["V-CLS", "V-QA", "V-RET", "V-MRET"]
|
| 22 |
TASKS_V = ['Video-Overall'] + SUB_TASKS_V + ALL_DATASETS_SPLITS['video']
|
| 23 |
COLUMN_NAMES_V = BASE_COLS + TASKS_V
|
| 24 |
DATA_TITLE_TYPE_V = BASE_DATA_TITLE_TYPE + \
|
| 25 |
+
['number'] * len(TASKS_V)
|
| 26 |
|
| 27 |
SUB_TASKS_D = ['ViDoRe-V1', 'ViDoRe-V2', 'VisRAG', 'VisDoc-OOD']
|
| 28 |
TASKS_D = ['Visdoc-Overall'] + SUB_TASKS_D + ALL_DATASETS_SPLITS['visdoc']
|
| 29 |
COLUMN_NAMES_D = BASE_COLS + TASKS_D
|
| 30 |
DATA_TITLE_TYPE_D = BASE_DATA_TITLE_TYPE + \
|
| 31 |
+
['number'] * len(TASKS_D)
|
| 32 |
|
| 33 |
TABLE_INTRODUCTION = """**MMEB**: Massive MultiModal Embedding Benchmark \n
|
| 34 |
Models are ranked based on **Overall**"""
|
|
|
|
| 41 |
Models are ranked based on **Video-Overall**"""
|
| 42 |
TABLE_INTRODUCTION_D = """**VisDoc**: Visual Document Understanding \n
|
| 43 |
Models are ranked based on **Visdoc-Overall** \n
|
| 44 |
+
**⚠️ Your Attention Please! We have fixed the error found in the eval code of ViDoSeek-page and MMLongBench-page (See this [GitHub issue](https://github.com/TIGER-AI-Lab/VLM2Vec/issues/167) for more info), and the two datasets' scores have been temporarily removed from the leaderboard. We have added two new columns to the leaderboard, ViDoSeek-page-fixed and MMLongBench-page-fixed. The scores of models submitted before --JANURARY 2026-- have been changed to ViDoSeek-page-old and MMLongBench-page-old, and will be replaced after a new submission. Here is the [list of models](https://huggingface.co/spaces/TIGER-Lab/MMEB-Leaderboard/blob/main/archive/page_old_scores.jsonl) that has to be fixed. We kindly ask the authors on this list to rerun their models on the visdoc datasets and re-submit a PR with your fixed scores. (Note: If you see a ⚠️ in your model's score, please rerun your model and re-submit your scores, and then your old scores will be placed a ✅ to indicate a fixed version.) We appreciate your cooperations!**"""
|
| 45 |
|
| 46 |
LEADERBOARD_INFO = """
|
| 47 |
## Dataset Summary
|
|
|
|
| 76 |
all_scores = {}
|
| 77 |
for modality, datasets_list in DATASETS.items(): # Ex.: ('image', {'I-CLS': [...], 'I-QA': [...]})
|
| 78 |
for sub_task, datasets in datasets_list.items(): # Ex.: ('I-CLS', ['VOC2007', 'N24News', ...])
|
| 79 |
+
# ========================= HARD CODED TEMPORARY FIX =================
|
| 80 |
+
if modality == 'visdoc' and sub_task == 'VisDoc-OOD':
|
| 81 |
+
datasets = datasets + ['ViDoSeek-page', 'MMLongBench-page']
|
| 82 |
+
# ====================================================================
|
| 83 |
for dataset in datasets: # Ex.: 'VOC2007'
|
| 84 |
score = raw_scores.get(modality, {}).get(dataset, 0.0)
|
| 85 |
score = 0.0 if isinstance(score, str) and "N/A" in score else score
|
|
|
|
| 177 |
|
| 178 |
return filtered_df[COLUMN_NAMES]
|
| 179 |
|
| 180 |
+
def save_ranking_summary(df, name, save_now=True, dir='rankings'):
|
| 181 |
csv_path, json_path = os.path.join(dir, f'{name}.csv'), os.path.join(dir, f'{name}.jsonl')
|
| 182 |
+
if save_now:
|
| 183 |
+
df.to_csv(csv_path, index=False)
|
| 184 |
+
df.to_json(json_path, orient='records', lines=True)
|
| 185 |
return csv_path, json_path
|
| 186 |
|
| 187 |
def download_ranking(df, name, format='csv', dir='rankings'):
|
| 188 |
+
csv_path, json_path = save_ranking_summary(df, name, save_now=False, dir=dir)
|
| 189 |
return csv_path if format == 'csv' else json_path
|