chaitnya26 zpn commited on
Commit
3a531b4
·
verified ·
0 Parent(s):

Duplicate from nomic-ai/colnomic-embed-multimodal-7b

Browse files

Co-authored-by: Zach Nussbaum <[email protected]>

.gitattributes ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,176 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model:
3
+ - Qwen/Qwen2.5-VL-7B-Instruct
4
+ library_name: peft
5
+ datasets:
6
+ - llamaindex/vdr-multilingual-train
7
+ - nomic-ai/colpali_train_set_split_by_source
8
+ language:
9
+ - en
10
+ - it
11
+ - fr
12
+ - de
13
+ - es
14
+ pipeline_tag: visual-document-retrieval
15
+ tags:
16
+ - vidore
17
+ - colpali
18
+ - multimodal_embedding
19
+ - multilingual_embedding
20
+ - Text-to-Visual Document (T→VD) retrieval
21
+ license: apache-2.0
22
+ ---
23
+
24
+ # ColNomic Embed Multimodal 7B: State-of-the-Art Visual Document Retrieval
25
+
26
+ `colnomic-embed-multimodal-7b` is a multi-vector state-of-the-art multimodal embedding model that excels at visual document retrieval tasks:
27
+
28
+ - **High Performance**: Achieves 62.7 NDCG@5 on Vidore-v2, outperforming all other models
29
+ - **Unified Text-Image Encoding**: Directly encodes interleaved text and images without complex preprocessing
30
+ - **Advanced Architecture**: 7B parameter multimodal embedding model
31
+ - **Fully Open-Source**: Model weights, training data, and code available
32
+
33
+ ## Performance
34
+
35
+ | Model | Avg. | ESG Restaurant Human | Econ Macro Multi. | AXA Multi. | MIT Bio | ESG Restaurant Synth. | ESG Restaurant Synth. Multi. | MIT Bio Multi. | AXA | Econ. Macro |
36
+ |-------|------|----------------------|-------------------|------------|---------|----------------------|----------------------------|---------------|-----|------------|
37
+ | **ColNomic Embed Multimodal 7B** | 62.7 | 73.9 | 54.7 | 61.3 | 66.1 | 57.3 | 56.7 | 64.2 | 68.3 | 61.6 |
38
+ | [ColNomic Embed Multimodal 3B](https://huggingface.co/nomic-ai/colnomic-embed-multimodal-3b) | 61.2 | 65.8 | 55.4 | 61.0 | 63.5 | 56.6 | 57.2 | 62.5 | 68.8 | 60.2 |
39
+ | T-Systems ColQwen2.5-3B | 59.9 | 72.1 | 51.2 | 60.0 | 65.3 | 51.7 | 53.3 | 61.7 | 69.3 | 54.8 |
40
+ | [Nomic Embed Multimodal 7B](https://huggingface.co/nomic-ai/nomic-embed-multimodal-7b) | 59.7 | 65.7 | 57.7 | 59.3 | 64.0 | 49.2 | 51.9 | 61.2 | 66.3 | 63.1 |
41
+ | GME Qwen2 7B | 59.0 | 65.8 | 56.2 | 55.4 | 64.0 | 54.3 | 56.7 | 55.1 | 60.7 | 62.9 |
42
+ | [Nomic Embed Multimodal 3B](https://huggingface.co/nomic-ai/nomic-embed-multimodal-3b) | 58.8 | 59.8 | 57.5 | 58.8 | 62.5 | 49.4 | 49.4 | 58.6 | 69.6 | 63.5 |
43
+ | Llama Index vdr-2b-multi-v1 | 58.4 | 63.1 | 52.8 | 61.0 | 60.6 | 50.3 | 51.2 | 56.9 | 68.8 | 61.2 |
44
+ | Voyage Multimodal 3 | 55.0 | 56.1 | 55.0 | 59.5 | 56.4 | 47.2 | 46.2 | 51.5 | 64.1 | 58.8 |
45
+
46
+ ## Getting Started
47
+
48
+ To use `colnomic-embed-multimodal-7b`, please install `colpali` from source
49
+
50
+ ```bash
51
+ pip install git+https://github.com/illuin-tech/colpali.git
52
+ ```
53
+
54
+
55
+ ```python
56
+ import torch
57
+ from PIL import Image
58
+ from transformers.utils.import_utils import is_flash_attn_2_available
59
+
60
+ from colpali_engine.models import ColQwen2_5, ColQwen2_5_Processor
61
+
62
+ model_name = "nomic-ai/colnomic-embed-multimodal-7b"
63
+
64
+ model = ColQwen2_5.from_pretrained(
65
+ model_name,
66
+ torch_dtype=torch.bfloat16,
67
+ device_map="cuda:0", # or "mps" if on Apple Silicon
68
+ attn_implementation="flash_attention_2" if is_flash_attn_2_available() else None,
69
+ ).eval()
70
+
71
+ processor = ColQwen2_5_Processor.from_pretrained(model_name)
72
+
73
+ # Your inputs
74
+ images = [
75
+ Image.new("RGB", (128, 128), color="white"),
76
+ Image.new("RGB", (64, 32), color="black"),
77
+ ]
78
+ queries = [
79
+ "What is the organizational structure for our R&D department?",
80
+ "Can you provide a breakdown of last year’s financial performance?",
81
+ ]
82
+
83
+ # Process the inputs
84
+ batch_images = processor.process_images(images).to(model.device)
85
+ batch_queries = processor.process_queries(queries).to(model.device)
86
+
87
+ # Forward pass
88
+ with torch.no_grad():
89
+ image_embeddings = model(**batch_images)
90
+ query_embeddings = model(**batch_queries)
91
+
92
+ scores = processor.score_multi_vector(query_embeddings, image_embeddings)
93
+ ```
94
+
95
+ ## Model Architecture
96
+
97
+ - **Total Parameters**: 7B
98
+ - **Training Approach**: Fine-tuned from Qwen2.5-VL 7B Instruct
99
+ - **Architecture Type**: Vision-Language Model with unified text and image input processing
100
+ - **Key Innovations**:
101
+ - Same-source sampling to create harder in-batch negatives
102
+ - Multi-vector output option for enhanced performance
103
+
104
+ ## Integration with RAG Workflows
105
+
106
+ Nomic Embed Multimodal 7B seamlessly integrates with Retrieval Augmented Generation (RAG) workflows:
107
+
108
+ 1. **Direct Document Embedding**: Skip OCR and complex processing by directly embedding document page images
109
+ 2. **Faster Processing**: Eliminate preprocessing steps for quicker indexing
110
+ 3. **More Complete Information**: Capture both textual and visual cues in a single embedding
111
+ 4. **Simple Implementation**: Use the same API for both text and images
112
+
113
+ ## Recommended Use Cases
114
+
115
+ The model excels at handling real-world document retrieval scenarios that challenge traditional text-only systems:
116
+
117
+ - **Research Papers**: Capture equations, diagrams, and tables
118
+ - **Technical Documentation**: Encode code blocks, flowcharts, and screenshots
119
+ - **Product Catalogs**: Represent images, specifications, and pricing tables
120
+ - **Financial Reports**: Embed charts, graphs, and numerical data
121
+ - **Visually Rich Content**: Where layout and visual information are important
122
+ - **Multilingual Documents**: Where visual context provides important cues
123
+
124
+ ## Training Details
125
+
126
+ ColNomic Embed Multimodal 7B was developed through several key innovations:
127
+
128
+ 1. **Sampling From the Same Source**: Forcing sampling from the same dataset source creates harder in-batch negatives, preventing the model from learning dataset artifacts.
129
+
130
+ 2. **Multi-Vector Configuration**: Providing a multi-vector variant that achieves higher performance than the dense variant.
131
+
132
+ ## Limitations
133
+
134
+ - Performance may vary when processing documents with unconventional layouts or unusual visual elements
135
+ - While it handles multiple languages, performance is strongest on English content
136
+ - Processing very large or complex documents may require dividing them into smaller chunks
137
+ - Performance on documents with handwriting or heavily stylized fonts may be reduced
138
+
139
+ ## Join the Nomic Community
140
+
141
+ - Nomic Embed Ecosystem: [https://www.nomic.ai/embed](https://www.nomic.ai/embed)
142
+ - Website: [https://nomic.ai](https://nomic.ai)
143
+ - Twitter: [https://twitter.com/nomic_ai](https://twitter.com/nomic_ai)
144
+ - Discord: [https://discord.gg/myY5YDR8z8](https://discord.gg/myY5YDR8z8)
145
+
146
+ ## Citation
147
+
148
+ If you find this model useful in your research or applications, please consider citing:
149
+
150
+ ```bibtex
151
+ @misc{faysse2024colpaliefficientdocumentretrieval,
152
+ title={ColPali: Efficient Document Retrieval with Vision Language Models},
153
+ author={Manuel Faysse and Hugues Sibille and Tony Wu and Bilel Omrani and Gautier Viaud and Céline Hudelot and Pierre Colombo},
154
+ year={2024},
155
+ eprint={2407.01449},
156
+ archivePrefix={arXiv},
157
+ primaryClass={cs.IR},
158
+ url={https://arxiv.org/abs/2407.01449},
159
+ }
160
+ @misc{ma2024unifyingmultimodalretrievaldocument,
161
+ title={Unifying Multimodal Retrieval via Document Screenshot Embedding},
162
+ author={Xueguang Ma and Sheng-Chieh Lin and Minghan Li and Wenhu Chen and Jimmy Lin},
163
+ year={2024},
164
+ eprint={2406.11251},
165
+ archivePrefix={arXiv},
166
+ primaryClass={cs.IR},
167
+ url={https://arxiv.org/abs/2406.11251},
168
+ }
169
+ @misc{nomicembedmultimodal2025,
170
+ title={Nomic Embed Multimodal: Interleaved Text, Image, and Screenshots for Visual Document Retrieval},
171
+ author={Nomic Team},
172
+ year={2025},
173
+ publisher={Nomic AI},
174
+ url={https://nomic.ai/blog/posts/nomic-embed-multimodal},
175
+ }
176
+ ```
adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "nomic-ai/colqwen2.5-7B-base",
5
+ "bias": "none",
6
+ "eva_config": null,
7
+ "exclude_modules": null,
8
+ "fan_in_fan_out": false,
9
+ "inference_mode": true,
10
+ "init_lora_weights": "gaussian",
11
+ "layer_replication": null,
12
+ "layers_pattern": null,
13
+ "layers_to_transform": null,
14
+ "loftq_config": {},
15
+ "lora_alpha": 32,
16
+ "lora_bias": false,
17
+ "lora_dropout": 0.1,
18
+ "megatron_config": null,
19
+ "megatron_core": "megatron.core",
20
+ "modules_to_save": null,
21
+ "peft_type": "LORA",
22
+ "r": 32,
23
+ "rank_pattern": {},
24
+ "revision": null,
25
+ "target_modules": "(.*(model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)",
26
+ "task_type": "FEATURE_EXTRACTION",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa6cba28368342ac6b235f49fcc3e0a5bd891d56df490c12bcbc6dc52c8377f3
3
+ size 323489536
added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
chat_template.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}"
3
+ }
git_hash.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ d3f780b24ba870601a046b4f35d28d827739446c
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": true,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.48145466,
8
+ 0.4578275,
9
+ 0.40821073
10
+ ],
11
+ "image_processor_type": "Qwen2VLImageProcessor",
12
+ "image_std": [
13
+ 0.26862954,
14
+ 0.26130258,
15
+ 0.27577711
16
+ ],
17
+ "max_pixels": 602112,
18
+ "merge_size": 2,
19
+ "min_pixels": 3136,
20
+ "patch_size": 14,
21
+ "processor_class": "ColQwen2_5_Processor",
22
+ "resample": 3,
23
+ "rescale_factor": 0.00392156862745098,
24
+ "size": {
25
+ "longest_edge": 12845056,
26
+ "shortest_edge": 3136
27
+ },
28
+ "temporal_patch_size": 2
29
+ }
processor_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_num_visual_tokens": 1280,
3
+ "processor_class": "ColQwen2_5_Processor"
4
+ }
results.json ADDED
@@ -0,0 +1,1104 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "vidore/arxivqa_test_subsampled":{
3
+ "ndcg_at_1":0.83,
4
+ "ndcg_at_3":0.87533,
5
+ "ndcg_at_5":0.88678,
6
+ "ndcg_at_10":0.89447,
7
+ "ndcg_at_20":0.89988,
8
+ "ndcg_at_50":0.90114,
9
+ "ndcg_at_100":0.90247,
10
+ "map_at_1":0.83,
11
+ "map_at_3":0.86467,
12
+ "map_at_5":0.87097,
13
+ "map_at_10":0.87409,
14
+ "map_at_20":0.87551,
15
+ "map_at_50":0.87574,
16
+ "map_at_100":0.87587,
17
+ "recall_at_1":0.83,
18
+ "recall_at_3":0.906,
19
+ "recall_at_5":0.934,
20
+ "recall_at_10":0.958,
21
+ "recall_at_20":0.98,
22
+ "recall_at_50":0.986,
23
+ "recall_at_100":0.994,
24
+ "precision_at_1":0.83,
25
+ "precision_at_3":0.302,
26
+ "precision_at_5":0.1868,
27
+ "precision_at_10":0.0958,
28
+ "precision_at_20":0.049,
29
+ "precision_at_50":0.01972,
30
+ "precision_at_100":0.00994,
31
+ "mrr_at_1":0.832,
32
+ "mrr_at_3":0.8663333333333333,
33
+ "mrr_at_5":0.8716333333333333,
34
+ "mrr_at_10":0.8752095238095239,
35
+ "mrr_at_20":0.8764481089945486,
36
+ "mrr_at_50":0.8769236116400513,
37
+ "mrr_at_100":0.8770087474728994,
38
+ "naucs_at_1_max":0.7799094444766933,
39
+ "naucs_at_1_std":-0.08580135833284933,
40
+ "naucs_at_1_diff1":0.9595692807801705,
41
+ "naucs_at_3_max":0.7955579394878525,
42
+ "naucs_at_3_std":-0.11155809841667234,
43
+ "naucs_at_3_diff1":0.9224725351133368,
44
+ "naucs_at_5_max":0.7705061822708856,
45
+ "naucs_at_5_std":-0.25026172084995735,
46
+ "naucs_at_5_diff1":0.9217666864725705,
47
+ "naucs_at_10_max":0.7739540260548654,
48
+ "naucs_at_10_std":-0.21324085189631833,
49
+ "naucs_at_10_diff1":0.9221910986616823,
50
+ "naucs_at_20_max":0.7708216619981358,
51
+ "naucs_at_20_std":-0.0996732026143794,
52
+ "naucs_at_20_diff1":0.8905228758169942,
53
+ "naucs_at_50_max":0.816593303988261,
54
+ "naucs_at_50_std":-0.19987995198079136,
55
+ "naucs_at_50_diff1":0.8832866479925231,
56
+ "naucs_at_100_max":0.9564270152505304,
57
+ "naucs_at_100_std":0.04310613134141329,
58
+ "naucs_at_100_diff1":0.9564270152505304
59
+ },
60
+ "vidore/docvqa_test_subsampled":{
61
+ "ndcg_at_1":0.5388,
62
+ "ndcg_at_3":0.59542,
63
+ "ndcg_at_5":0.61269,
64
+ "ndcg_at_10":0.62919,
65
+ "ndcg_at_20":0.64259,
66
+ "ndcg_at_50":0.64929,
67
+ "ndcg_at_100":0.65683,
68
+ "map_at_1":0.5388,
69
+ "map_at_3":0.5813,
70
+ "map_at_5":0.59084,
71
+ "map_at_10":0.59764,
72
+ "map_at_20":0.6013,
73
+ "map_at_50":0.60241,
74
+ "map_at_100":0.60307,
75
+ "recall_at_1":0.5388,
76
+ "recall_at_3":0.63636,
77
+ "recall_at_5":0.67849,
78
+ "recall_at_10":0.72949,
79
+ "recall_at_20":0.78271,
80
+ "recall_at_50":0.81596,
81
+ "recall_at_100":0.86253,
82
+ "precision_at_1":0.5388,
83
+ "precision_at_3":0.21212,
84
+ "precision_at_5":0.1357,
85
+ "precision_at_10":0.07295,
86
+ "precision_at_20":0.03914,
87
+ "precision_at_50":0.01632,
88
+ "precision_at_100":0.00863,
89
+ "mrr_at_1":0.5432372505543237,
90
+ "mrr_at_3":0.5824094604582412,
91
+ "mrr_at_5":0.5903917220990392,
92
+ "mrr_at_10":0.5980097138633723,
93
+ "mrr_at_20":0.601820515422298,
94
+ "mrr_at_50":0.6029928998761527,
95
+ "mrr_at_100":0.6035667141572244,
96
+ "naucs_at_1_max":0.8430735352775856,
97
+ "naucs_at_1_std":0.5939128313043085,
98
+ "naucs_at_1_diff1":0.9018027917306838,
99
+ "naucs_at_3_max":0.8808230124613547,
100
+ "naucs_at_3_std":0.712681445860455,
101
+ "naucs_at_3_diff1":0.8646220134947054,
102
+ "naucs_at_5_max":0.8912698205476521,
103
+ "naucs_at_5_std":0.7525159249929689,
104
+ "naucs_at_5_diff1":0.8414611325660849,
105
+ "naucs_at_10_max":0.8852509111743405,
106
+ "naucs_at_10_std":0.817204422380503,
107
+ "naucs_at_10_diff1":0.8264628195305581,
108
+ "naucs_at_20_max":0.8512017211379178,
109
+ "naucs_at_20_std":0.8215729110215341,
110
+ "naucs_at_20_diff1":0.7992889048812711,
111
+ "naucs_at_50_max":0.8402551716801772,
112
+ "naucs_at_50_std":0.8733758748271155,
113
+ "naucs_at_50_diff1":0.7960833079573262,
114
+ "naucs_at_100_max":0.838262046357413,
115
+ "naucs_at_100_std":0.904184740422419,
116
+ "naucs_at_100_diff1":0.7996865824491111
117
+ },
118
+ "vidore/infovqa_test_subsampled":{
119
+ "ndcg_at_1":0.90081,
120
+ "ndcg_at_3":0.92934,
121
+ "ndcg_at_5":0.93352,
122
+ "ndcg_at_10":0.93669,
123
+ "ndcg_at_20":0.93871,
124
+ "ndcg_at_50":0.9403,
125
+ "ndcg_at_100":0.94099,
126
+ "map_at_1":0.90081,
127
+ "map_at_3":0.9224,
128
+ "map_at_5":0.92473,
129
+ "map_at_10":0.92598,
130
+ "map_at_20":0.92652,
131
+ "map_at_50":0.92677,
132
+ "map_at_100":0.92684,
133
+ "recall_at_1":0.90081,
134
+ "recall_at_3":0.94939,
135
+ "recall_at_5":0.95951,
136
+ "recall_at_10":0.96964,
137
+ "recall_at_20":0.97773,
138
+ "recall_at_50":0.98583,
139
+ "recall_at_100":0.98988,
140
+ "precision_at_1":0.90081,
141
+ "precision_at_3":0.31646,
142
+ "precision_at_5":0.1919,
143
+ "precision_at_10":0.09696,
144
+ "precision_at_20":0.04889,
145
+ "precision_at_50":0.01972,
146
+ "precision_at_100":0.0099,
147
+ "mrr_at_1":0.9008097165991903,
148
+ "mrr_at_3":0.9210526315789472,
149
+ "mrr_at_5":0.9248987854251012,
150
+ "mrr_at_10":0.9258040935672512,
151
+ "mrr_at_20":0.926239353314252,
152
+ "mrr_at_50":0.9265337558342984,
153
+ "mrr_at_100":0.9266429029872886,
154
+ "naucs_at_1_max":0.666687407456454,
155
+ "naucs_at_1_std":0.04348917328850024,
156
+ "naucs_at_1_diff1":0.9460185154989313,
157
+ "naucs_at_3_max":0.887692938235415,
158
+ "naucs_at_3_std":0.3724699632820944,
159
+ "naucs_at_3_diff1":0.9582075346992748,
160
+ "naucs_at_5_max":0.9121353932948266,
161
+ "naucs_at_5_std":0.46274982003896187,
162
+ "naucs_at_5_diff1":0.9608195637805694,
163
+ "naucs_at_10_max":0.9553640600031791,
164
+ "naucs_at_10_std":0.6527520622928011,
165
+ "naucs_at_10_diff1":0.9564661819784096,
166
+ "naucs_at_20_max":0.9643814216187027,
167
+ "naucs_at_20_std":0.8342696288850537,
168
+ "naucs_at_20_diff1":0.9643814216187027,
169
+ "naucs_at_50_max":1.0,
170
+ "naucs_at_50_std":0.960323608891707,
171
+ "naucs_at_50_diff1":0.9626852988386237,
172
+ "naucs_at_100_max":1.0,
173
+ "naucs_at_100_std":1.0,
174
+ "naucs_at_100_diff1":0.9477594183740937
175
+ },
176
+ "vidore/tabfquad_test_subsampled":{
177
+ "ndcg_at_1":0.925,
178
+ "ndcg_at_3":0.95561,
179
+ "ndcg_at_5":0.95991,
180
+ "ndcg_at_10":0.96464,
181
+ "ndcg_at_20":0.96464,
182
+ "ndcg_at_50":0.96464,
183
+ "ndcg_at_100":0.96464,
184
+ "map_at_1":0.925,
185
+ "map_at_3":0.94881,
186
+ "map_at_5":0.95113,
187
+ "map_at_10":0.95314,
188
+ "map_at_20":0.95314,
189
+ "map_at_50":0.95314,
190
+ "map_at_100":0.95314,
191
+ "recall_at_1":0.925,
192
+ "recall_at_3":0.975,
193
+ "recall_at_5":0.98571,
194
+ "recall_at_10":1.0,
195
+ "recall_at_20":1.0,
196
+ "recall_at_50":1.0,
197
+ "recall_at_100":1.0,
198
+ "precision_at_1":0.925,
199
+ "precision_at_3":0.325,
200
+ "precision_at_5":0.19714,
201
+ "precision_at_10":0.1,
202
+ "precision_at_20":0.05,
203
+ "precision_at_50":0.02,
204
+ "precision_at_100":0.01,
205
+ "mrr_at_1":0.9214285714285714,
206
+ "mrr_at_3":0.9464285714285714,
207
+ "mrr_at_5":0.9487499999999999,
208
+ "mrr_at_10":0.9507624716553289,
209
+ "mrr_at_20":0.9507624716553289,
210
+ "mrr_at_50":0.9507624716553289,
211
+ "mrr_at_100":0.9507624716553289,
212
+ "naucs_at_1_max":0.5344804588502072,
213
+ "naucs_at_1_std":0.26817393624116476,
214
+ "naucs_at_1_diff1":0.9383086567960518,
215
+ "naucs_at_3_max":0.9416433239962675,
216
+ "naucs_at_3_std":0.8989595838335351,
217
+ "naucs_at_3_diff1":0.98132586367881,
218
+ "naucs_at_5_max":0.967320261437913,
219
+ "naucs_at_5_std":0.8558590102707857,
220
+ "naucs_at_5_diff1":0.967320261437913,
221
+ "naucs_at_10_max":1.0,
222
+ "naucs_at_10_std":1.0,
223
+ "naucs_at_10_diff1":1.0,
224
+ "naucs_at_20_max":1.0,
225
+ "naucs_at_20_std":1.0,
226
+ "naucs_at_20_diff1":1.0,
227
+ "naucs_at_50_max":1.0,
228
+ "naucs_at_50_std":1.0,
229
+ "naucs_at_50_diff1":1.0,
230
+ "naucs_at_100_max":1.0,
231
+ "naucs_at_100_std":1.0,
232
+ "naucs_at_100_diff1":1.0
233
+ },
234
+ "vidore/tatdqa_test":{
235
+ "ndcg_at_1":0.70109,
236
+ "ndcg_at_3":0.79187,
237
+ "ndcg_at_5":0.81111,
238
+ "ndcg_at_10":0.82342,
239
+ "ndcg_at_20":0.8274,
240
+ "ndcg_at_50":0.83026,
241
+ "ndcg_at_100":0.83293,
242
+ "map_at_1":0.70109,
243
+ "map_at_3":0.77005,
244
+ "map_at_5":0.78071,
245
+ "map_at_10":0.78598,
246
+ "map_at_20":0.78706,
247
+ "map_at_50":0.78751,
248
+ "map_at_100":0.78775,
249
+ "recall_at_1":0.70109,
250
+ "recall_at_3":0.8548,
251
+ "recall_at_5":0.90158,
252
+ "recall_at_10":0.93864,
253
+ "recall_at_20":0.95443,
254
+ "recall_at_50":0.96902,
255
+ "recall_at_100":0.98542,
256
+ "precision_at_1":0.70109,
257
+ "precision_at_3":0.28493,
258
+ "precision_at_5":0.18032,
259
+ "precision_at_10":0.09386,
260
+ "precision_at_20":0.04772,
261
+ "precision_at_50":0.01938,
262
+ "precision_at_100":0.00985,
263
+ "mrr_at_1":0.7023086269744836,
264
+ "mrr_at_3":0.7714661806399354,
265
+ "mrr_at_5":0.782978938841637,
266
+ "mrr_at_10":0.7879385330478899,
267
+ "mrr_at_20":0.7889867733618092,
268
+ "mrr_at_50":0.7893770561483674,
269
+ "mrr_at_100":0.7896057919808999,
270
+ "naucs_at_1_max":0.25184895001768326,
271
+ "naucs_at_1_std":0.037017257538968965,
272
+ "naucs_at_1_diff1":0.8508843501202716,
273
+ "naucs_at_3_max":0.28836101906742934,
274
+ "naucs_at_3_std":0.0895587034833779,
275
+ "naucs_at_3_diff1":0.7677154118662606,
276
+ "naucs_at_5_max":0.33141107758129335,
277
+ "naucs_at_5_std":0.1402343443047445,
278
+ "naucs_at_5_diff1":0.726113129605695,
279
+ "naucs_at_10_max":0.3753484131980775,
280
+ "naucs_at_10_std":0.21181536222478634,
281
+ "naucs_at_10_diff1":0.7110698774619854,
282
+ "naucs_at_20_max":0.398127644970406,
283
+ "naucs_at_20_std":0.2900002928945751,
284
+ "naucs_at_20_diff1":0.6944425419244158,
285
+ "naucs_at_50_max":0.4094042090297014,
286
+ "naucs_at_50_std":0.3282254760208566,
287
+ "naucs_at_50_diff1":0.665997592036148,
288
+ "naucs_at_100_max":0.2601633836764814,
289
+ "naucs_at_100_std":0.2532897060083396,
290
+ "naucs_at_100_diff1":0.6808982214080673
291
+ },
292
+ "vidore/shiftproject_test":{
293
+ "ndcg_at_1":0.84,
294
+ "ndcg_at_3":0.91809,
295
+ "ndcg_at_5":0.91809,
296
+ "ndcg_at_10":0.92499,
297
+ "ndcg_at_20":0.92499,
298
+ "ndcg_at_50":0.92499,
299
+ "ndcg_at_100":0.92673,
300
+ "map_at_1":0.84,
301
+ "map_at_3":0.9,
302
+ "map_at_5":0.9,
303
+ "map_at_10":0.9031,
304
+ "map_at_20":0.9031,
305
+ "map_at_50":0.9031,
306
+ "map_at_100":0.90328,
307
+ "recall_at_1":0.84,
308
+ "recall_at_3":0.97,
309
+ "recall_at_5":0.97,
310
+ "recall_at_10":0.99,
311
+ "recall_at_20":0.99,
312
+ "recall_at_50":0.99,
313
+ "recall_at_100":1.0,
314
+ "precision_at_1":0.84,
315
+ "precision_at_3":0.32333,
316
+ "precision_at_5":0.194,
317
+ "precision_at_10":0.099,
318
+ "precision_at_20":0.0495,
319
+ "precision_at_50":0.0198,
320
+ "precision_at_100":0.01,
321
+ "mrr_at_1":0.84,
322
+ "mrr_at_3":0.8983333333333333,
323
+ "mrr_at_5":0.8983333333333333,
324
+ "mrr_at_10":0.9014285714285714,
325
+ "mrr_at_20":0.9014285714285714,
326
+ "mrr_at_50":0.9014285714285714,
327
+ "mrr_at_100":0.9016208791208791,
328
+ "naucs_at_1_max":0.1685076092292596,
329
+ "naucs_at_1_std":-0.2145925380461453,
330
+ "naucs_at_1_diff1":0.7751595483554242,
331
+ "naucs_at_3_max":0.6498599439775861,
332
+ "naucs_at_3_std":0.36834733893557625,
333
+ "naucs_at_3_diff1":0.664021164021167,
334
+ "naucs_at_5_max":0.6498599439775937,
335
+ "naucs_at_5_std":0.36834733893557176,
336
+ "naucs_at_5_diff1":0.6640211640211615,
337
+ "naucs_at_10_max":0.7222222222222276,
338
+ "naucs_at_10_std":0.5541549953314738,
339
+ "naucs_at_10_diff1":0.8692810457516413,
340
+ "naucs_at_20_max":0.7222222222222276,
341
+ "naucs_at_20_std":0.5541549953314738,
342
+ "naucs_at_20_diff1":0.8692810457516413,
343
+ "naucs_at_50_max":0.7222222222222041,
344
+ "naucs_at_50_std":0.554154995331464,
345
+ "naucs_at_50_diff1":0.8692810457516374,
346
+ "naucs_at_100_max":null,
347
+ "naucs_at_100_std":null,
348
+ "naucs_at_100_diff1":null
349
+ },
350
+ "vidore/syntheticDocQA_artificial_intelligence_test":{
351
+ "ndcg_at_1":0.98,
352
+ "ndcg_at_3":0.99262,
353
+ "ndcg_at_5":0.99262,
354
+ "ndcg_at_10":0.99262,
355
+ "ndcg_at_20":0.99262,
356
+ "ndcg_at_50":0.99262,
357
+ "ndcg_at_100":0.99262,
358
+ "map_at_1":0.98,
359
+ "map_at_3":0.99,
360
+ "map_at_5":0.99,
361
+ "map_at_10":0.99,
362
+ "map_at_20":0.99,
363
+ "map_at_50":0.99,
364
+ "map_at_100":0.99,
365
+ "recall_at_1":0.98,
366
+ "recall_at_3":1.0,
367
+ "recall_at_5":1.0,
368
+ "recall_at_10":1.0,
369
+ "recall_at_20":1.0,
370
+ "recall_at_50":1.0,
371
+ "recall_at_100":1.0,
372
+ "precision_at_1":0.98,
373
+ "precision_at_3":0.33333,
374
+ "precision_at_5":0.2,
375
+ "precision_at_10":0.1,
376
+ "precision_at_20":0.05,
377
+ "precision_at_50":0.02,
378
+ "precision_at_100":0.01,
379
+ "mrr_at_1":0.99,
380
+ "mrr_at_3":0.995,
381
+ "mrr_at_5":0.995,
382
+ "mrr_at_10":0.995,
383
+ "mrr_at_20":0.995,
384
+ "mrr_at_50":0.995,
385
+ "mrr_at_100":0.995,
386
+ "naucs_at_1_max":0.540149393090569,
387
+ "naucs_at_1_std":-0.6909430438842186,
388
+ "naucs_at_1_diff1":1.0,
389
+ "naucs_at_3_max":1.0,
390
+ "naucs_at_3_std":1.0,
391
+ "naucs_at_3_diff1":1.0,
392
+ "naucs_at_5_max":1.0,
393
+ "naucs_at_5_std":1.0,
394
+ "naucs_at_5_diff1":1.0,
395
+ "naucs_at_10_max":1.0,
396
+ "naucs_at_10_std":1.0,
397
+ "naucs_at_10_diff1":1.0,
398
+ "naucs_at_20_max":1.0,
399
+ "naucs_at_20_std":1.0,
400
+ "naucs_at_20_diff1":1.0,
401
+ "naucs_at_50_max":null,
402
+ "naucs_at_50_std":null,
403
+ "naucs_at_50_diff1":null,
404
+ "naucs_at_100_max":null,
405
+ "naucs_at_100_std":null,
406
+ "naucs_at_100_diff1":null
407
+ },
408
+ "vidore/syntheticDocQA_energy_test":{
409
+ "ndcg_at_1":0.95,
410
+ "ndcg_at_3":0.95631,
411
+ "ndcg_at_5":0.96492,
412
+ "ndcg_at_10":0.96781,
413
+ "ndcg_at_20":0.96781,
414
+ "ndcg_at_50":0.96976,
415
+ "ndcg_at_100":0.96976,
416
+ "map_at_1":0.95,
417
+ "map_at_3":0.955,
418
+ "map_at_5":0.96,
419
+ "map_at_10":0.961,
420
+ "map_at_20":0.961,
421
+ "map_at_50":0.96129,
422
+ "map_at_100":0.96129,
423
+ "recall_at_1":0.95,
424
+ "recall_at_3":0.96,
425
+ "recall_at_5":0.98,
426
+ "recall_at_10":0.99,
427
+ "recall_at_20":0.99,
428
+ "recall_at_50":1.0,
429
+ "recall_at_100":1.0,
430
+ "precision_at_1":0.95,
431
+ "precision_at_3":0.32,
432
+ "precision_at_5":0.196,
433
+ "precision_at_10":0.099,
434
+ "precision_at_20":0.0495,
435
+ "precision_at_50":0.02,
436
+ "precision_at_100":0.01,
437
+ "mrr_at_1":0.95,
438
+ "mrr_at_3":0.955,
439
+ "mrr_at_5":0.96,
440
+ "mrr_at_10":0.961,
441
+ "mrr_at_20":0.961,
442
+ "mrr_at_50":0.9613030303030302,
443
+ "mrr_at_100":0.9613030303030302,
444
+ "naucs_at_1_max":0.32362278244631343,
445
+ "naucs_at_1_std":-0.7713352007469629,
446
+ "naucs_at_1_diff1":1.0,
447
+ "naucs_at_3_max":0.8395191409897237,
448
+ "naucs_at_3_std":-0.6762371615312763,
449
+ "naucs_at_3_diff1":1.0,
450
+ "naucs_at_5_max":0.6790382819794609,
451
+ "naucs_at_5_std":-1.445845004668519,
452
+ "naucs_at_5_diff1":1.0,
453
+ "naucs_at_10_max":1.0,
454
+ "naucs_at_10_std":-1.1517273576097316,
455
+ "naucs_at_10_diff1":1.0,
456
+ "naucs_at_20_max":1.0,
457
+ "naucs_at_20_std":-1.1517273576097316,
458
+ "naucs_at_20_diff1":1.0,
459
+ "naucs_at_50_max":null,
460
+ "naucs_at_50_std":null,
461
+ "naucs_at_50_diff1":null,
462
+ "naucs_at_100_max":null,
463
+ "naucs_at_100_std":null,
464
+ "naucs_at_100_diff1":null
465
+ },
466
+ "vidore/syntheticDocQA_government_reports_test":{
467
+ "ndcg_at_1":0.91,
468
+ "ndcg_at_3":0.94655,
469
+ "ndcg_at_5":0.95085,
470
+ "ndcg_at_10":0.95743,
471
+ "ndcg_at_20":0.95743,
472
+ "ndcg_at_50":0.95743,
473
+ "ndcg_at_100":0.95743,
474
+ "map_at_1":0.91,
475
+ "map_at_3":0.93833,
476
+ "map_at_5":0.94083,
477
+ "map_at_10":0.94361,
478
+ "map_at_20":0.94361,
479
+ "map_at_50":0.94361,
480
+ "map_at_100":0.94361,
481
+ "recall_at_1":0.91,
482
+ "recall_at_3":0.97,
483
+ "recall_at_5":0.98,
484
+ "recall_at_10":1.0,
485
+ "recall_at_20":1.0,
486
+ "recall_at_50":1.0,
487
+ "recall_at_100":1.0,
488
+ "precision_at_1":0.91,
489
+ "precision_at_3":0.32333,
490
+ "precision_at_5":0.196,
491
+ "precision_at_10":0.1,
492
+ "precision_at_20":0.05,
493
+ "precision_at_50":0.02,
494
+ "precision_at_100":0.01,
495
+ "mrr_at_1":0.92,
496
+ "mrr_at_3":0.9433333333333332,
497
+ "mrr_at_5":0.9483333333333333,
498
+ "mrr_at_10":0.9494444444444444,
499
+ "mrr_at_20":0.9494444444444444,
500
+ "mrr_at_50":0.9494444444444444,
501
+ "mrr_at_100":0.9494444444444444,
502
+ "naucs_at_1_max":0.6774561676522453,
503
+ "naucs_at_1_std":0.14721444133208894,
504
+ "naucs_at_1_diff1":0.9564270152505436,
505
+ "naucs_at_3_max":0.8513849984438244,
506
+ "naucs_at_3_std":-0.24929971988795643,
507
+ "naucs_at_3_diff1":0.9564270152505466,
508
+ "naucs_at_5_max":1.0,
509
+ "naucs_at_5_std":-0.43534080298785716,
510
+ "naucs_at_5_diff1":0.9346405228758136,
511
+ "naucs_at_10_max":1.0,
512
+ "naucs_at_10_std":1.0,
513
+ "naucs_at_10_diff1":1.0,
514
+ "naucs_at_20_max":1.0,
515
+ "naucs_at_20_std":1.0,
516
+ "naucs_at_20_diff1":1.0,
517
+ "naucs_at_50_max":null,
518
+ "naucs_at_50_std":null,
519
+ "naucs_at_50_diff1":null,
520
+ "naucs_at_100_max":null,
521
+ "naucs_at_100_std":null,
522
+ "naucs_at_100_diff1":null
523
+ },
524
+ "vidore/syntheticDocQA_healthcare_industry_test":{
525
+ "ndcg_at_1":0.98,
526
+ "ndcg_at_3":0.99262,
527
+ "ndcg_at_5":0.99262,
528
+ "ndcg_at_10":0.99262,
529
+ "ndcg_at_20":0.99262,
530
+ "ndcg_at_50":0.99262,
531
+ "ndcg_at_100":0.99262,
532
+ "map_at_1":0.98,
533
+ "map_at_3":0.99,
534
+ "map_at_5":0.99,
535
+ "map_at_10":0.99,
536
+ "map_at_20":0.99,
537
+ "map_at_50":0.99,
538
+ "map_at_100":0.99,
539
+ "recall_at_1":0.98,
540
+ "recall_at_3":1.0,
541
+ "recall_at_5":1.0,
542
+ "recall_at_10":1.0,
543
+ "recall_at_20":1.0,
544
+ "recall_at_50":1.0,
545
+ "recall_at_100":1.0,
546
+ "precision_at_1":0.98,
547
+ "precision_at_3":0.33333,
548
+ "precision_at_5":0.2,
549
+ "precision_at_10":0.1,
550
+ "precision_at_20":0.05,
551
+ "precision_at_50":0.02,
552
+ "precision_at_100":0.01,
553
+ "mrr_at_1":0.98,
554
+ "mrr_at_3":0.99,
555
+ "mrr_at_5":0.99,
556
+ "mrr_at_10":0.99,
557
+ "mrr_at_20":0.99,
558
+ "mrr_at_50":0.99,
559
+ "mrr_at_100":0.99,
560
+ "naucs_at_1_max":0.7222222222222248,
561
+ "naucs_at_1_std":-0.5634920634920563,
562
+ "naucs_at_1_diff1":1.0,
563
+ "naucs_at_3_max":1.0,
564
+ "naucs_at_3_std":1.0,
565
+ "naucs_at_3_diff1":1.0,
566
+ "naucs_at_5_max":1.0,
567
+ "naucs_at_5_std":1.0,
568
+ "naucs_at_5_diff1":1.0,
569
+ "naucs_at_10_max":1.0,
570
+ "naucs_at_10_std":1.0,
571
+ "naucs_at_10_diff1":1.0,
572
+ "naucs_at_20_max":1.0,
573
+ "naucs_at_20_std":1.0,
574
+ "naucs_at_20_diff1":1.0,
575
+ "naucs_at_50_max":null,
576
+ "naucs_at_50_std":null,
577
+ "naucs_at_50_diff1":null,
578
+ "naucs_at_100_max":null,
579
+ "naucs_at_100_std":null,
580
+ "naucs_at_100_diff1":null
581
+ },
582
+ "vidore/synthetic_mit_biomedical_tissue_interactions_unfiltered":{
583
+ "ndcg_at_1":0.66875,
584
+ "ndcg_at_3":0.64833,
585
+ "ndcg_at_5":0.66136,
586
+ "ndcg_at_10":0.7002,
587
+ "ndcg_at_20":0.72343,
588
+ "ndcg_at_50":0.74221,
589
+ "ndcg_at_100":0.7504,
590
+ "map_at_1":0.41001,
591
+ "map_at_3":0.53585,
592
+ "map_at_5":0.57453,
593
+ "map_at_10":0.61456,
594
+ "map_at_20":0.63039,
595
+ "map_at_50":0.63734,
596
+ "map_at_100":0.63926,
597
+ "recall_at_1":0.41001,
598
+ "recall_at_3":0.59207,
599
+ "recall_at_5":0.66951,
600
+ "recall_at_10":0.7842,
601
+ "recall_at_20":0.84612,
602
+ "recall_at_50":0.90015,
603
+ "recall_at_100":0.92697,
604
+ "precision_at_1":0.66875,
605
+ "precision_at_3":0.39375,
606
+ "precision_at_5":0.29125,
607
+ "precision_at_10":0.18938,
608
+ "precision_at_20":0.11,
609
+ "precision_at_50":0.05087,
610
+ "precision_at_100":0.0275,
611
+ "mrr_at_1":0.65,
612
+ "mrr_at_3":0.7364583333333331,
613
+ "mrr_at_5":0.7480208333333331,
614
+ "mrr_at_10":0.7537127976190475,
615
+ "mrr_at_20":0.7549048402255638,
616
+ "mrr_at_50":0.755591370159045,
617
+ "mrr_at_100":0.755591370159045,
618
+ "naucs_at_1_max":0.3941383204398901,
619
+ "naucs_at_1_std":-0.03446929320007161,
620
+ "naucs_at_1_diff1":0.4503277010269023,
621
+ "naucs_at_3_max":0.029009734517562568,
622
+ "naucs_at_3_std":0.005402681474374929,
623
+ "naucs_at_3_diff1":-0.048706505991403734,
624
+ "naucs_at_5_max":-0.08207834225628742,
625
+ "naucs_at_5_std":-0.0032984001708723115,
626
+ "naucs_at_5_diff1":-0.17283336776461594,
627
+ "naucs_at_10_max":-0.17977285443640584,
628
+ "naucs_at_10_std":-0.04541684253923944,
629
+ "naucs_at_10_diff1":-0.23295672062273604,
630
+ "naucs_at_20_max":-0.170593905587224,
631
+ "naucs_at_20_std":0.015384406511413186,
632
+ "naucs_at_20_diff1":-0.2551053459454954,
633
+ "naucs_at_50_max":-0.22984435384481694,
634
+ "naucs_at_50_std":0.04443493398633732,
635
+ "naucs_at_50_diff1":-0.27937793023113594,
636
+ "naucs_at_100_max":-0.25957527630817795,
637
+ "naucs_at_100_std":0.057081217133354005,
638
+ "naucs_at_100_diff1":-0.3033742727110607
639
+ },
640
+ "vidore/synthetic_economics_macro_economy_2024_filtered_v1.0":{
641
+ "ndcg_at_1":0.65517,
642
+ "ndcg_at_3":0.62816,
643
+ "ndcg_at_5":0.6159,
644
+ "ndcg_at_10":0.58912,
645
+ "ndcg_at_20":0.61706,
646
+ "ndcg_at_50":0.67765,
647
+ "ndcg_at_100":0.71471,
648
+ "map_at_1":0.08673,
649
+ "map_at_3":0.19894,
650
+ "map_at_5":0.26716,
651
+ "map_at_10":0.33292,
652
+ "map_at_20":0.40089,
653
+ "map_at_50":0.45667,
654
+ "map_at_100":0.4844,
655
+ "recall_at_1":0.08673,
656
+ "recall_at_3":0.24596,
657
+ "recall_at_5":0.34869,
658
+ "recall_at_10":0.4683,
659
+ "recall_at_20":0.62193,
660
+ "recall_at_50":0.80907,
661
+ "recall_at_100":0.92119,
662
+ "precision_at_1":0.65517,
663
+ "precision_at_3":0.58621,
664
+ "precision_at_5":0.55172,
665
+ "precision_at_10":0.42931,
666
+ "precision_at_20":0.32672,
667
+ "precision_at_50":0.19759,
668
+ "precision_at_100":0.1281,
669
+ "mrr_at_1":0.6896551724137931,
670
+ "mrr_at_3":0.7902298850574713,
671
+ "mrr_at_5":0.7971264367816091,
672
+ "mrr_at_10":0.8015051997810618,
673
+ "mrr_at_20":0.8015051997810618,
674
+ "mrr_at_50":0.8023262178434591,
675
+ "mrr_at_100":0.8023262178434591,
676
+ "naucs_at_1_max":-0.046307070115366736,
677
+ "naucs_at_1_std":0.15790688174002224,
678
+ "naucs_at_1_diff1":-0.14987305341556462,
679
+ "naucs_at_3_max":0.11671339502888234,
680
+ "naucs_at_3_std":0.11834362034275059,
681
+ "naucs_at_3_diff1":-0.17762652230301346,
682
+ "naucs_at_5_max":0.10679449683484428,
683
+ "naucs_at_5_std":0.09453878709557821,
684
+ "naucs_at_5_diff1":-0.1792193824994377,
685
+ "naucs_at_10_max":0.10651036963192122,
686
+ "naucs_at_10_std":0.207447480653671,
687
+ "naucs_at_10_diff1":-0.23752907816911517,
688
+ "naucs_at_20_max":0.05980777985340397,
689
+ "naucs_at_20_std":0.23365933479100473,
690
+ "naucs_at_20_diff1":-0.2025898492844735,
691
+ "naucs_at_50_max":-0.012623885278852142,
692
+ "naucs_at_50_std":0.2463485456131626,
693
+ "naucs_at_50_diff1":-0.18909262376408734,
694
+ "naucs_at_100_max":-0.08872391317584416,
695
+ "naucs_at_100_std":0.17948141514389135,
696
+ "naucs_at_100_diff1":-0.186930857132979
697
+ },
698
+ "vidore/synthetic_rse_restaurant_filtered_v1.0":{
699
+ "ndcg_at_1":0.52632,
700
+ "ndcg_at_3":0.52504,
701
+ "ndcg_at_5":0.57289,
702
+ "ndcg_at_10":0.62003,
703
+ "ndcg_at_20":0.65505,
704
+ "ndcg_at_50":0.68979,
705
+ "ndcg_at_100":0.70204,
706
+ "map_at_1":0.29474,
707
+ "map_at_3":0.40396,
708
+ "map_at_5":0.45972,
709
+ "map_at_10":0.50866,
710
+ "map_at_20":0.5276,
711
+ "map_at_50":0.54534,
712
+ "map_at_100":0.55161,
713
+ "recall_at_1":0.29474,
714
+ "recall_at_3":0.48191,
715
+ "recall_at_5":0.61543,
716
+ "recall_at_10":0.75002,
717
+ "recall_at_20":0.85537,
718
+ "recall_at_50":0.94841,
719
+ "recall_at_100":0.98246,
720
+ "precision_at_1":0.52632,
721
+ "precision_at_3":0.34503,
722
+ "precision_at_5":0.28421,
723
+ "precision_at_10":0.19649,
724
+ "precision_at_20":0.12281,
725
+ "precision_at_50":0.06912,
726
+ "precision_at_100":0.03877,
727
+ "mrr_at_1":0.543859649122807,
728
+ "mrr_at_3":0.6403508771929824,
729
+ "mrr_at_5":0.6526315789473685,
730
+ "mrr_at_10":0.6602548036758563,
731
+ "mrr_at_20":0.6651739829371409,
732
+ "mrr_at_50":0.6655248601301235,
733
+ "mrr_at_100":0.6655248601301235,
734
+ "naucs_at_1_max":0.07429395157545386,
735
+ "naucs_at_1_std":0.03015644215746205,
736
+ "naucs_at_1_diff1":0.029188519373247488,
737
+ "naucs_at_3_max":-0.14556834354246306,
738
+ "naucs_at_3_std":-0.03551874034674307,
739
+ "naucs_at_3_diff1":-0.03426575331185958,
740
+ "naucs_at_5_max":-0.132667608690031,
741
+ "naucs_at_5_std":0.046883073270420496,
742
+ "naucs_at_5_diff1":0.020789980948042533,
743
+ "naucs_at_10_max":-0.15350074281217774,
744
+ "naucs_at_10_std":0.021602996730365333,
745
+ "naucs_at_10_diff1":0.10610855445931551,
746
+ "naucs_at_20_max":-0.2870595843942612,
747
+ "naucs_at_20_std":-0.053200969005437526,
748
+ "naucs_at_20_diff1":0.009614869287472011,
749
+ "naucs_at_50_max":-0.3747840510110352,
750
+ "naucs_at_50_std":-0.1315278407645466,
751
+ "naucs_at_50_diff1":-0.08682413153737878,
752
+ "naucs_at_100_max":-0.3797677839881161,
753
+ "naucs_at_100_std":-0.11609328262961503,
754
+ "naucs_at_100_diff1":-0.12296867865157128
755
+ },
756
+ "vidore/synthetic_axa_filtered_v1.0":{
757
+ "ndcg_at_1":0.61111,
758
+ "ndcg_at_3":0.6462,
759
+ "ndcg_at_5":0.68292,
760
+ "ndcg_at_10":0.70017,
761
+ "ndcg_at_20":0.72201,
762
+ "ndcg_at_50":0.7613,
763
+ "ndcg_at_100":0.7658,
764
+ "map_at_1":0.29544,
765
+ "map_at_3":0.41914,
766
+ "map_at_5":0.49958,
767
+ "map_at_10":0.57784,
768
+ "map_at_20":0.61104,
769
+ "map_at_50":0.63144,
770
+ "map_at_100":0.63315,
771
+ "recall_at_1":0.29544,
772
+ "recall_at_3":0.51973,
773
+ "recall_at_5":0.66378,
774
+ "recall_at_10":0.78472,
775
+ "recall_at_20":0.85945,
776
+ "recall_at_50":0.97412,
777
+ "recall_at_100":0.98765,
778
+ "precision_at_1":0.61111,
779
+ "precision_at_3":0.46296,
780
+ "precision_at_5":0.42222,
781
+ "precision_at_10":0.31111,
782
+ "precision_at_20":0.19167,
783
+ "precision_at_50":0.09111,
784
+ "precision_at_100":0.04667,
785
+ "mrr_at_1":0.5555555555555556,
786
+ "mrr_at_3":0.6759259259259259,
787
+ "mrr_at_5":0.7037037037037037,
788
+ "mrr_at_10":0.7037037037037037,
789
+ "mrr_at_20":0.7037037037037037,
790
+ "mrr_at_50":0.7060185185185185,
791
+ "mrr_at_100":0.7060185185185185,
792
+ "naucs_at_1_max":0.06370640145678692,
793
+ "naucs_at_1_std":0.42153041887902504,
794
+ "naucs_at_1_diff1":0.35531111328419196,
795
+ "naucs_at_3_max":-0.49137089517804067,
796
+ "naucs_at_3_std":-0.41095639069482337,
797
+ "naucs_at_3_diff1":0.01707919337550571,
798
+ "naucs_at_5_max":-0.6120658234932744,
799
+ "naucs_at_5_std":-0.474389941591449,
800
+ "naucs_at_5_diff1":-0.165061956148533,
801
+ "naucs_at_10_max":-0.6883907404624792,
802
+ "naucs_at_10_std":-0.45513671453258775,
803
+ "naucs_at_10_diff1":-0.42316346500786917,
804
+ "naucs_at_20_max":-0.733957515102311,
805
+ "naucs_at_20_std":-0.4723325528538845,
806
+ "naucs_at_20_diff1":-0.4807537601912433,
807
+ "naucs_at_50_max":-0.7013881037720148,
808
+ "naucs_at_50_std":-0.45324759826825234,
809
+ "naucs_at_50_diff1":-0.4882986634795409,
810
+ "naucs_at_100_max":-0.6924248438890106,
811
+ "naucs_at_100_std":-0.4400487565870318,
812
+ "naucs_at_100_diff1":-0.4661845036852439
813
+ },
814
+ "vidore/synthetic_rse_restaurant_filtered_v1.0_multilingual":{
815
+ "ndcg_at_1":0.54825,
816
+ "ndcg_at_3":0.53008,
817
+ "ndcg_at_5":0.56705,
818
+ "ndcg_at_10":0.6257,
819
+ "ndcg_at_20":0.6625,
820
+ "ndcg_at_50":0.69353,
821
+ "ndcg_at_100":0.70465,
822
+ "map_at_1":0.29816,
823
+ "map_at_3":0.4016,
824
+ "map_at_5":0.4527,
825
+ "map_at_10":0.50738,
826
+ "map_at_20":0.52962,
827
+ "map_at_50":0.5466,
828
+ "map_at_100":0.55224,
829
+ "recall_at_1":0.29816,
830
+ "recall_at_3":0.48014,
831
+ "recall_at_5":0.59266,
832
+ "recall_at_10":0.76588,
833
+ "recall_at_20":0.87318,
834
+ "recall_at_50":0.95235,
835
+ "recall_at_100":0.98465,
836
+ "precision_at_1":0.54825,
837
+ "precision_at_3":0.34795,
838
+ "precision_at_5":0.28158,
839
+ "precision_at_10":0.19956,
840
+ "precision_at_20":0.12654,
841
+ "precision_at_50":0.06991,
842
+ "precision_at_100":0.03877,
843
+ "mrr_at_1":0.5394736842105263,
844
+ "mrr_at_3":0.6381578947368421,
845
+ "mrr_at_5":0.6528508771929824,
846
+ "mrr_at_10":0.666278543581175,
847
+ "mrr_at_20":0.6693822460076329,
848
+ "mrr_at_50":0.6696454039023697,
849
+ "mrr_at_100":0.6697038834345334,
850
+ "naucs_at_1_max":-0.10947777032499051,
851
+ "naucs_at_1_std":-0.08674361998375714,
852
+ "naucs_at_1_diff1":0.057249370353071344,
853
+ "naucs_at_3_max":-0.06687224875668692,
854
+ "naucs_at_3_std":0.0484364629589946,
855
+ "naucs_at_3_diff1":-0.05462426569089335,
856
+ "naucs_at_5_max":-0.08403937224497868,
857
+ "naucs_at_5_std":0.07293419772827552,
858
+ "naucs_at_5_diff1":-0.05958666027516894,
859
+ "naucs_at_10_max":-0.10838130867204888,
860
+ "naucs_at_10_std":0.05587982504285361,
861
+ "naucs_at_10_diff1":0.002192750308197299,
862
+ "naucs_at_20_max":-0.1628596976822229,
863
+ "naucs_at_20_std":0.06415536794747155,
864
+ "naucs_at_20_diff1":-0.0901131230050924,
865
+ "naucs_at_50_max":-0.2471444987505313,
866
+ "naucs_at_50_std":0.02178464235724526,
867
+ "naucs_at_50_diff1":-0.16484319783599452,
868
+ "naucs_at_100_max":-0.2476326734790779,
869
+ "naucs_at_100_std":0.04921366491713892,
870
+ "naucs_at_100_diff1":-0.1925485969942229
871
+ },
872
+ "vidore/synthetic_axa_filtered_v1.0_multilingual":{
873
+ "ndcg_at_1":0.56944,
874
+ "ndcg_at_3":0.60114,
875
+ "ndcg_at_5":0.61308,
876
+ "ndcg_at_10":0.63301,
877
+ "ndcg_at_20":0.65962,
878
+ "ndcg_at_50":0.70087,
879
+ "ndcg_at_100":0.71794,
880
+ "map_at_1":0.26807,
881
+ "map_at_3":0.38274,
882
+ "map_at_5":0.44543,
883
+ "map_at_10":0.51783,
884
+ "map_at_20":0.55091,
885
+ "map_at_50":0.57108,
886
+ "map_at_100":0.57464,
887
+ "recall_at_1":0.26807,
888
+ "recall_at_3":0.47345,
889
+ "recall_at_5":0.56537,
890
+ "recall_at_10":0.69703,
891
+ "recall_at_20":0.78343,
892
+ "recall_at_50":0.90152,
893
+ "recall_at_100":0.97889,
894
+ "precision_at_1":0.56944,
895
+ "precision_at_3":0.43056,
896
+ "precision_at_5":0.38056,
897
+ "precision_at_10":0.27917,
898
+ "precision_at_20":0.17569,
899
+ "precision_at_50":0.08583,
900
+ "precision_at_100":0.04569,
901
+ "mrr_at_1":0.5,
902
+ "mrr_at_3":0.6481481481481483,
903
+ "mrr_at_5":0.6550925925925927,
904
+ "mrr_at_10":0.6566358024691359,
905
+ "mrr_at_20":0.6577041785375121,
906
+ "mrr_at_50":0.659896412080552,
907
+ "mrr_at_100":0.6607831288544433,
908
+ "naucs_at_1_max":0.16468725740231643,
909
+ "naucs_at_1_std":0.37919017176717934,
910
+ "naucs_at_1_diff1":0.30214037198755495,
911
+ "naucs_at_3_max":-0.2419012234949494,
912
+ "naucs_at_3_std":-0.11880659562104584,
913
+ "naucs_at_3_diff1":-0.05983117053027332,
914
+ "naucs_at_5_max":-0.3435266789444122,
915
+ "naucs_at_5_std":-0.15122373891433766,
916
+ "naucs_at_5_diff1":-0.14301810021483025,
917
+ "naucs_at_10_max":-0.4049299119799346,
918
+ "naucs_at_10_std":-0.11551373909975436,
919
+ "naucs_at_10_diff1":-0.3087348082925975,
920
+ "naucs_at_20_max":-0.4557080881366001,
921
+ "naucs_at_20_std":-0.1316035057344777,
922
+ "naucs_at_20_diff1":-0.3767740930238455,
923
+ "naucs_at_50_max":-0.48221867587284106,
924
+ "naucs_at_50_std":-0.16520884486012927,
925
+ "naucs_at_50_diff1":-0.437539802211166,
926
+ "naucs_at_100_max":-0.5041035001503106,
927
+ "naucs_at_100_std":-0.19246332353391876,
928
+ "naucs_at_100_diff1":-0.450877397124322
929
+ },
930
+ "vidore/synthetic_mit_biomedical_tissue_interactions_unfiltered_multilingual":{
931
+ "ndcg_at_1":0.62656,
932
+ "ndcg_at_3":0.62898,
933
+ "ndcg_at_5":0.64249,
934
+ "ndcg_at_10":0.67441,
935
+ "ndcg_at_20":0.69935,
936
+ "ndcg_at_50":0.72034,
937
+ "ndcg_at_100":0.7304,
938
+ "map_at_1":0.37553,
939
+ "map_at_3":0.51341,
940
+ "map_at_5":0.55134,
941
+ "map_at_10":0.58654,
942
+ "map_at_20":0.60275,
943
+ "map_at_50":0.61047,
944
+ "map_at_100":0.61272,
945
+ "recall_at_1":0.37553,
946
+ "recall_at_3":0.58317,
947
+ "recall_at_5":0.6651,
948
+ "recall_at_10":0.76267,
949
+ "recall_at_20":0.83047,
950
+ "recall_at_50":0.89402,
951
+ "recall_at_100":0.93005,
952
+ "precision_at_1":0.62656,
953
+ "precision_at_3":0.39115,
954
+ "precision_at_5":0.28812,
955
+ "precision_at_10":0.18453,
956
+ "precision_at_20":0.10859,
957
+ "precision_at_50":0.05059,
958
+ "precision_at_100":0.02756,
959
+ "mrr_at_1":0.6203125,
960
+ "mrr_at_3":0.7114583333333329,
961
+ "mrr_at_5":0.7251302083333327,
962
+ "mrr_at_10":0.7312574404761898,
963
+ "mrr_at_20":0.7330658025942703,
964
+ "mrr_at_50":0.7337640283895484,
965
+ "mrr_at_100":0.733871254289439,
966
+ "naucs_at_1_max":0.2039250412963779,
967
+ "naucs_at_1_std":-0.07024884868861188,
968
+ "naucs_at_1_diff1":0.45606770903022215,
969
+ "naucs_at_3_max":0.058212985355640734,
970
+ "naucs_at_3_std":-0.05959626421557797,
971
+ "naucs_at_3_diff1":-0.029625222250297182,
972
+ "naucs_at_5_max":0.012523889018321476,
973
+ "naucs_at_5_std":-0.05950139173785889,
974
+ "naucs_at_5_diff1":-0.12093229795276991,
975
+ "naucs_at_10_max":-0.07885246896142604,
976
+ "naucs_at_10_std":-0.07584435766192119,
977
+ "naucs_at_10_diff1":-0.19868179893511964,
978
+ "naucs_at_20_max":-0.08104765093094657,
979
+ "naucs_at_20_std":-0.020633911923048417,
980
+ "naucs_at_20_diff1":-0.2329474014068329,
981
+ "naucs_at_50_max":-0.10191575025539294,
982
+ "naucs_at_50_std":0.006497205696773289,
983
+ "naucs_at_50_diff1":-0.2654796386435107,
984
+ "naucs_at_100_max":-0.12912675673565918,
985
+ "naucs_at_100_std":0.002658413394124888,
986
+ "naucs_at_100_diff1":-0.286740024806076
987
+ },
988
+ "vidore/synthetics_economics_macro_economy_2024_filtered_v1.0_multilingual":{
989
+ "ndcg_at_1":0.56897,
990
+ "ndcg_at_3":0.55347,
991
+ "ndcg_at_5":0.54697,
992
+ "ndcg_at_10":0.54353,
993
+ "ndcg_at_20":0.5754,
994
+ "ndcg_at_50":0.64435,
995
+ "ndcg_at_100":0.68217,
996
+ "map_at_1":0.06995,
997
+ "map_at_3":0.16788,
998
+ "map_at_5":0.22262,
999
+ "map_at_10":0.29539,
1000
+ "map_at_20":0.35806,
1001
+ "map_at_50":0.41596,
1002
+ "map_at_100":0.44377,
1003
+ "recall_at_1":0.06995,
1004
+ "recall_at_3":0.22001,
1005
+ "recall_at_5":0.31105,
1006
+ "recall_at_10":0.45156,
1007
+ "recall_at_20":0.60587,
1008
+ "recall_at_50":0.80629,
1009
+ "recall_at_100":0.91617,
1010
+ "precision_at_1":0.56897,
1011
+ "precision_at_3":0.52155,
1012
+ "precision_at_5":0.49483,
1013
+ "precision_at_10":0.4069,
1014
+ "precision_at_20":0.31207,
1015
+ "precision_at_50":0.19595,
1016
+ "precision_at_100":0.1275,
1017
+ "mrr_at_1":0.5689655172413793,
1018
+ "mrr_at_3":0.7040229885057472,
1019
+ "mrr_at_5":0.7173850574712642,
1020
+ "mrr_at_10":0.7212370005473451,
1021
+ "mrr_at_20":0.7221635921046082,
1022
+ "mrr_at_50":0.7223688466202075,
1023
+ "mrr_at_100":0.7223688466202075,
1024
+ "naucs_at_1_max":0.008287845664171043,
1025
+ "naucs_at_1_std":0.1276905868842376,
1026
+ "naucs_at_1_diff1":0.11685328065061831,
1027
+ "naucs_at_3_max":0.04231917732527594,
1028
+ "naucs_at_3_std":0.09097968258247123,
1029
+ "naucs_at_3_diff1":0.07039779015197889,
1030
+ "naucs_at_5_max":0.02450452954672108,
1031
+ "naucs_at_5_std":0.08382631941812721,
1032
+ "naucs_at_5_diff1":0.04903893807286711,
1033
+ "naucs_at_10_max":0.06258885863723426,
1034
+ "naucs_at_10_std":0.1489118063390922,
1035
+ "naucs_at_10_diff1":-0.001772551183085018,
1036
+ "naucs_at_20_max":0.05127122134169797,
1037
+ "naucs_at_20_std":0.16207987797849627,
1038
+ "naucs_at_20_diff1":-0.037084808056486104,
1039
+ "naucs_at_50_max":0.012079926402609917,
1040
+ "naucs_at_50_std":0.16924631618333094,
1041
+ "naucs_at_50_diff1":-0.09727554327201936,
1042
+ "naucs_at_100_max":-0.03658022842524374,
1043
+ "naucs_at_100_std":0.12849977057146159,
1044
+ "naucs_at_100_diff1":-0.10217040873502983
1045
+ },
1046
+ "vidore/restaurant_esg_reports_beir":{
1047
+ "ndcg_at_1":0.71795,
1048
+ "ndcg_at_3":0.71188,
1049
+ "ndcg_at_5":0.73949,
1050
+ "ndcg_at_10":0.77205,
1051
+ "ndcg_at_20":0.78712,
1052
+ "ndcg_at_50":0.80866,
1053
+ "ndcg_at_100":0.81008,
1054
+ "map_at_1":0.50513,
1055
+ "map_at_3":0.62382,
1056
+ "map_at_5":0.6711,
1057
+ "map_at_10":0.70078,
1058
+ "map_at_20":0.7096,
1059
+ "map_at_50":0.71767,
1060
+ "map_at_100":0.71825,
1061
+ "recall_at_1":0.50513,
1062
+ "recall_at_3":0.68668,
1063
+ "recall_at_5":0.76906,
1064
+ "recall_at_10":0.85835,
1065
+ "recall_at_20":0.90725,
1066
+ "recall_at_50":0.97476,
1067
+ "recall_at_100":0.98071,
1068
+ "precision_at_1":0.73077,
1069
+ "precision_at_3":0.38462,
1070
+ "precision_at_5":0.28846,
1071
+ "precision_at_10":0.175,
1072
+ "precision_at_20":0.09712,
1073
+ "precision_at_50":0.045,
1074
+ "precision_at_100":0.02288,
1075
+ "mrr_at_1":0.75,
1076
+ "mrr_at_3":0.8076923076923077,
1077
+ "mrr_at_5":0.8125,
1078
+ "mrr_at_10":0.8226495726495725,
1079
+ "mrr_at_20":0.8226495726495725,
1080
+ "mrr_at_50":0.8233363858363858,
1081
+ "mrr_at_100":0.8233363858363858,
1082
+ "naucs_at_1_max":0.012222350607946745,
1083
+ "naucs_at_1_std":-0.09466371539982948,
1084
+ "naucs_at_1_diff1":0.5022751223737485,
1085
+ "naucs_at_3_max":-0.03148341001971903,
1086
+ "naucs_at_3_std":-0.017366247697995007,
1087
+ "naucs_at_3_diff1":-0.2806074619839212,
1088
+ "naucs_at_5_max":-0.1368097107443235,
1089
+ "naucs_at_5_std":-0.05686309477532154,
1090
+ "naucs_at_5_diff1":-0.34542841772396893,
1091
+ "naucs_at_10_max":-0.11047704445523453,
1092
+ "naucs_at_10_std":0.016760999608655385,
1093
+ "naucs_at_10_diff1":-0.36219386821062893,
1094
+ "naucs_at_20_max":-0.18118046303544919,
1095
+ "naucs_at_20_std":0.00660518472055942,
1096
+ "naucs_at_20_diff1":-0.35322674833881423,
1097
+ "naucs_at_50_max":-0.15496181424390146,
1098
+ "naucs_at_50_std":0.025351601144475155,
1099
+ "naucs_at_50_diff1":-0.3271145822876798,
1100
+ "naucs_at_100_max":-0.14539299033899847,
1101
+ "naucs_at_100_std":0.020343129158802924,
1102
+ "naucs_at_100_diff1":-0.3237772572696072
1103
+ }
1104
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5eee858c5123a4279c3e1f7b81247343f356ac767940b2692a928ad929543214
3
+ size 11422063
tokenizer_config.json ADDED
@@ -0,0 +1,210 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0]['role'] == 'system' %}\n {{- messages[0]['content'] }}\n {%- else %}\n {{- 'You are a helpful assistant.' }}\n {%- endif %}\n {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0]['role'] == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n {%- else %}\n {{- '<|im_start|>system\\nYou are a helpful assistant.<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {{- '<|im_start|>' + message.role }}\n {%- if message.content %}\n {{- '\\n' + message.content }}\n {%- endif %}\n {%- for tool_call in message.tool_calls %}\n {%- if tool_call.function is defined %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '\\n<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {{- tool_call.arguments | tojson }}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
199
+ "clean_up_tokenization_spaces": false,
200
+ "eos_token": "<|im_end|>",
201
+ "errors": "replace",
202
+ "extra_special_tokens": {},
203
+ "max_num_visual_tokens": 1280,
204
+ "model_max_length": 131072,
205
+ "pad_token": "<|endoftext|>",
206
+ "processor_class": "ColQwen2_5_Processor",
207
+ "split_special_tokens": false,
208
+ "tokenizer_class": "Qwen2Tokenizer",
209
+ "unk_token": null
210
+ }
training_config.yml ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config:
2
+ (): colpali_engine.trainer.colmodel_training.ColModelTrainingConfig
3
+ output_dir: !path ../../../models/colqwen2_5_train_single_source_3ep_r32_512bs_vdr_7b_1e4
4
+ processor:
5
+ (): colpali_engine.utils.transformers_wrappers.AllPurposeWrapper
6
+ class_to_instanciate: !ext colpali_engine.models.ColQwen2_5_Processor
7
+ pretrained_model_name_or_path: "Qwen/Qwen2.5-VL-7B-Instruct" # "./models/paligemma-3b-mix-448"
8
+ max_num_visual_tokens: 1280
9
+ # num_image_tokens: 2048
10
+ # max_length: 50
11
+
12
+ model:
13
+ (): colpali_engine.utils.transformers_wrappers.AllPurposeWrapper
14
+ class_to_instanciate: !ext colpali_engine.models.ColQwen2_5
15
+ pretrained_model_name_or_path: "Qwen/Qwen2.5-VL-7B-Instruct"
16
+ torch_dtype: !ext torch.bfloat16
17
+ use_cache: false
18
+ attn_implementation: "flash_attention_2"
19
+ # device_map: "auto"
20
+ # quantization_config:
21
+ # (): transformers.BitsAndBytesConfig
22
+ # load_in_4bit: true
23
+ # bnb_4bit_quant_type: "nf4"
24
+ # bnb_4bit_compute_dtype: "bfloat16"
25
+ # bnb_4bit_use_double_quant: true
26
+
27
+ dataset_loading_func: !ext colpali_engine.utils.dataset_transformation.load_train_set_colpali_vdr
28
+ eval_dataset_loader: !import ../data/test_data.yaml
29
+
30
+ # max_length: 50
31
+ run_eval: true
32
+ loss_func:
33
+ (): colpali_engine.loss.late_interaction_losses.ColbertPairwiseCELoss
34
+ tr_args:
35
+ (): transformers.training_args.TrainingArguments
36
+ output_dir: null
37
+ overwrite_output_dir: true
38
+ num_train_epochs: 3
39
+ per_device_train_batch_size: 512
40
+ ddp_find_unused_parameters: false
41
+ gradient_checkpointing: true
42
+ gradient_checkpointing_kwargs: { "use_reentrant": false }
43
+ # gradient_checkpointing: true
44
+ # 6 x 8 gpus = 48 batch size
45
+ # gradient_accumulation_steps: 4
46
+ per_device_eval_batch_size: 16
47
+ eval_strategy: "no"
48
+ dataloader_num_workers: 0
49
+ # bf16: true
50
+ accelerator_config:
51
+ split_batches: true
52
+ save_steps: 500
53
+ logging_steps: 10
54
+ eval_steps: 100
55
+ warmup_ratio: 0.01
56
+ learning_rate: 1e-4
57
+ save_total_limit: 1
58
+ # resume_from_checkpoint: true
59
+ optim: "paged_adamw_8bit"
60
+ # wandb logging
61
+ # wandb_project: "mllm"
62
+ run_name: "colqwen2_5_train_single_source_3ep_r32_512bs_vdr_7b_1e4"
63
+ report_to: "wandb"
64
+
65
+
66
+ peft_config:
67
+ (): peft.LoraConfig
68
+ r: 32
69
+ lora_alpha: 32
70
+ lora_dropout: 0.1
71
+ init_lora_weights: "gaussian"
72
+ bias: "none"
73
+ task_type: "FEATURE_EXTRACTION"
74
+ target_modules: '(.*(model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)'
75
+ # target_modules: '(.*(language_model).*(down_proj|gate_proj|up_proj|k_proj|q_proj|v_proj|o_proj).*$|.*(custom_text_proj).*$)'
vocab.json ADDED
The diff for this file is too large to render. See raw diff