add AIBOM

Dear llamaindex,
We are a group of researchers investigating the usefulness of sharing AIBOMs (Artificial Intelligence Bill of Materials) to document AI models and to improve transparency in AI model supply chains. AIBOMs are machine-readable, structured inventories of components—such as datasets and models—used in the development of AI-powered systems.

We would like to emphasize that we have no financial or competing interests related to AIBOMs. Our sole interest is to advance the collective understanding of AIBOMs within both academia and industry. As part of this effort, we are contributing to randomly selected open and popular models on Hugging Face (like yours) and are happy to offer support to you and the maintainers of your model if needed.

Based on your model card (and some configuration information available in Hugging Face), we generated the AIBOM according to the CyclonDX (v1.6) standard (see https://cyclonedx.org/docs/1.6/json/). This AIBOM is generated as a JSON file by using the following open-source supporting tool: https://github.com/MSR4SBOM/ALOHA (technical details are available in the research paper: https://github.com/MSR4SBOM/ALOHA/blob/main/ALOHA.pdf). This tool is freely available online and can be downloaded and used at your own convenience. We are also happy to assist you directly if you need help generating or reviewing an AIBOM for your model.

The JSON file in this pull request is your AIBOM (see https://github.com/MSR4SBOM/ALOHA/blob/main/documentation.json for details on its structure). Clearly, the submitted AIBOM matches the current model information, yet it can be easily regenerated when the model evolves, using the aforementioned AIBOM generation tool.

We understand that initiatives like ours may raise questions, especially in open communities like Hugging Face. Therefore, we would like to further remark that our interest in AIBOMs is only to enhance the body of knowledge on AIBOMs and to make this easy and low-friction for maintainers of AI models and developers of AI-powered systems.

We open this pull request containing an AIBOM of your AI model, and hope it will be considered. We would also like to hear your opinion on the usefulness (or not) of AIBOM by answering a 3-minute anonymous survey: https://forms.gle/WGffSQD5dLoWttEe7.

Thanks in advance, and regards,
Riccardo D’Avino, Fatima Ahmed, Sabato Nocera, Simone Romano, Giuseppe Scanniello (University of Salerno, Italy),
Massimiliano Di Penta (University of Sannio, Italy),
The MSR4SBOM team

llamaindex_vdr-2b-multi-v1.json

@@ -0,0 +1,142 @@
+{
+    "bomFormat": "CycloneDX",
+    "specVersion": "1.6",
+    "serialNumber": "urn:uuid:00ca3e26-e155-40c6-801d-90be620fd041",
+    "version": 1,
+    "metadata": {
+        "timestamp": "2025-07-14T11:11:18.744037+00:00",
+        "component": {
+            "type": "machine-learning-model",
+            "bom-ref": "llamaindex/vdr-2b-multi-v1-9a12af53-e3c0-557a-86ff-8b132947222a",
+            "name": "llamaindex/vdr-2b-multi-v1",
+            "externalReferences": [
+                {
+                    "url": "https://huggingface.co/llamaindex/vdr-2b-multi-v1",
+                    "type": "documentation"
+                }
+            ],
+            "modelCard": {
+                "modelParameters": {
+                    "task": "image-to-text",
+                    "architectureFamily": "qwen2_vl",
+                    "modelArchitecture": "Qwen2VLForConditionalGeneration",
+                    "datasets": [
+                        {
+                            "ref": "llamaindex/vdr-multilingual-train-a46edb90-89ea-5b4e-a4aa-7bbe0726c573"
+                        }
+                    ]
+                },
+                "properties": [
+                    {
+                        "name": "library_name",
+                        "value": "sentence-transformers"
+                    },
+                    {
+                        "name": "base_model",
+                        "value": "MrLight/dse-qwen2-2b-mrl-v1"
+                    }
+                ]
+            },
+            "authors": [
+                {
+                    "name": "llamaindex"
+                }
+            ],
+            "licenses": [
+                {
+                    "license": {
+                        "id": "Apache-2.0",
+                        "url": "https://spdx.org/licenses/Apache-2.0.html"
+                    }
+                }
+            ],
+            "tags": [
+                "sentence-transformers",
+                "safetensors",
+                "qwen2_vl",
+                "image-to-text",
+                "transformers",
+                "Qwen2-VL",
+                "en",
+                "it",
+                "fr",
+                "de",
+                "es",
+                "dataset:llamaindex/vdr-multilingual-train",
+                "arxiv:2406.11251",
+                "base_model:MrLight/dse-qwen2-2b-mrl-v1",
+                "base_model:finetune:MrLight/dse-qwen2-2b-mrl-v1",
+                "license:apache-2.0",
+                "text-generation-inference",
+                "endpoints_compatible",
+                "region:us"
+            ]
+        }
+    },
+    "components": [
+        {
+            "type": "data",
+            "bom-ref": "llamaindex/vdr-multilingual-train-a46edb90-89ea-5b4e-a4aa-7bbe0726c573",
+            "name": "llamaindex/vdr-multilingual-train",
+            "data": [
+                {
+                    "type": "dataset",
+                    "bom-ref": "llamaindex/vdr-multilingual-train-a46edb90-89ea-5b4e-a4aa-7bbe0726c573",
+                    "name": "llamaindex/vdr-multilingual-train",
+                    "contents": {
+                        "url": "https://huggingface.co/datasets/llamaindex/vdr-multilingual-train",
+                        "properties": [
+                            {
+                                "name": "language",
+                                "value": "de, it, fr, es, en"
+                            },
+                            {
+                                "name": "size_categories",
+                                "value": "100K<n<1M"
+                            },
+                            {
+                                "name": "pretty_name",
+                                "value": "Multilingual Visual Document Retrieval"
+                            },
+                            {
+                                "name": "configs",
+                                "value": "Name of the dataset subset: en {\"split\": \"train\", \"path\": \"en/train-*\"}"
+                            },
+                            {
+                                "name": "configs",
+                                "value": "Name of the dataset subset: it {\"split\": \"train\", \"path\": \"it/train-*\"}"
+                            },
+                            {
+                                "name": "configs",
+                                "value": "Name of the dataset subset: fr {\"split\": \"train\", \"path\": \"fr/train-*\"}"
+                            },
+                            {
+                                "name": "configs",
+                                "value": "Name of the dataset subset: es {\"split\": \"train\", \"path\": \"es/train-*\"}"
+                            },
+                            {
+                                "name": "configs",
+                                "value": "Name of the dataset subset: de {\"split\": \"train\", \"path\": \"de/train-*\"}"
+                            },
+                            {
+                                "name": "license",
+                                "value": "apache-2.0"
+                            }
+                        ]
+                    },
+                    "governance": {
+                        "owners": [
+                            {
+                                "organization": {
+                                    "name": "llamaindex",
+                                    "url": "https://huggingface.co/llamaindex"
+                                }
+                            }
+                        ]
+                    },
+                    "description": "\n\t\n\t\t\n\t\tMultilingual Visual Document Retrieval Dataset\n\t\n\n\n\nThis dataset consists of 500k multilingual query image samples, collected and generated from scratch using public internet pdfs. The queries are synthetic and generated using VLMs (gemini-1.5-pro and Qwen2-VL-72B).\n\nIt was used to train the vdr-2b-multi-v1 retrieval multimodal, multilingual embedding model.\n\n\t\n\t\t\n\t\n\t\n\t\tHow it was created\n\t\n\nThis is the entire data pipeline used to create the Italian subset of this dataset. Each step\u2026 See the full description on the dataset page: https://huggingface.co/datasets/llamaindex/vdr-multilingual-train."
+                }
+            ]
+        }
+    ]
+}