{
    "bomFormat": "CycloneDX",
    "specVersion": "1.6",
    "serialNumber": "urn:uuid:f3d4b532-6fad-4b9a-b4cb-24dc7ba6e74a",
    "version": 1,
    "metadata": {
        "timestamp": "2025-06-05T09:39:11.687113+00:00",
        "component": {
            "type": "machine-learning-model",
            "bom-ref": "OpenGVLab/InternVL3-78B-e9d921a6-e013-5632-9136-d28da087616e",
            "name": "OpenGVLab/InternVL3-78B",
            "externalReferences": [
                {
                    "url": "https://huggingface.co/OpenGVLab/InternVL3-78B",
                    "type": "documentation"
                }
            ],
            "modelCard": {
                "modelParameters": {
                    "task": "image-text-to-text",
                    "architectureFamily": "internvl_chat",
                    "modelArchitecture": "InternVLChatModel",
                    "datasets": [
                        {
                            "ref": "OpenGVLab/MMPR-v1.2-f5ad7f01-75b1-5539-aff3-747fe24b14f6"
                        }
                    ]
                },
                "properties": [
                    {
                        "name": "library_name",
                        "value": "transformers"
                    },
                    {
                        "name": "base_model",
                        "value": "OpenGVLab/InternVL3-78B-Instruct"
                    },
                    {
                        "name": "base_model_relation",
                        "value": "finetune"
                    }
                ]
            },
            "authors": [
                {
                    "name": "OpenGVLab"
                }
            ],
            "licenses": [
                {
                    "license": {
                        "name": "qwen",
                        "url": "https://huggingface.co/Qwen/Qwen2.5-72B-Instruct/blob/main/LICENSE"
                    }
                }
            ],
            "description": "We introduce InternVL3, an advanced multimodal large language model (MLLM) series that demonstrates superior overall performance.Compared to InternVL 2.5, InternVL3 exhibits superior multimodal perception and reasoning capabilities, while further extending its multimodal capabilities to encompass tool usage, GUI agents, industrial image analysis, 3D vision perception, and more.Additionally, we compare InternVL3 with  Qwen2.5 Chat models, whose corresponding pre-trained base models are employed as the initialization of the langauge component in InternVL3. Benefitting from Native Multimodal Pre-Training, the InternVL3 series achieves even better overall text performance than the Qwen2.5 series.![image/png](https://huggingface.co/datasets/Weiyun1025/InternVL-Performance/resolve/main/internvl3/overall.png)",
            "tags": [
                "transformers",
                "safetensors",
                "internvl_chat",
                "feature-extraction",
                "internvl",
                "custom_code",
                "image-text-to-text",
                "conversational",
                "multilingual",
                "dataset:OpenGVLab/MMPR-v1.2",
                "arxiv:2312.14238",
                "arxiv:2404.16821",
                "arxiv:2412.05271",
                "arxiv:2411.10442",
                "arxiv:2504.10479",
                "arxiv:2412.09616",
                "base_model:OpenGVLab/InternVL3-78B-Instruct",
                "base_model:finetune:OpenGVLab/InternVL3-78B-Instruct",
                "license:other",
                "region:us"
            ]
        }
    },
    "components": [
        {
            "type": "data",
            "bom-ref": "OpenGVLab/MMPR-v1.2-f5ad7f01-75b1-5539-aff3-747fe24b14f6",
            "name": "OpenGVLab/MMPR-v1.2",
            "data": [
                {
                    "type": "dataset",
                    "bom-ref": "OpenGVLab/MMPR-v1.2-f5ad7f01-75b1-5539-aff3-747fe24b14f6",
                    "name": "OpenGVLab/MMPR-v1.2",
                    "contents": {
                        "url": "https://huggingface.co/datasets/OpenGVLab/MMPR-v1.2",
                        "properties": [
                            {
                                "name": "task_categories",
                                "value": "visual-question-answering"
                            },
                            {
                                "name": "language",
                                "value": "en"
                            },
                            {
                                "name": "size_categories",
                                "value": "1M<n<10M"
                            },
                            {
                                "name": "pretty_name",
                                "value": "MMPR-v1.2"
                            },
                            {
                                "name": "configs",
                                "value": "Name of the dataset subset: default {\"split\": \"train\", \"path\": \"annotations.zip\"}"
                            },
                            {
                                "name": "license",
                                "value": "mit"
                            }
                        ]
                    },
                    "governance": {
                        "owners": [
                            {
                                "organization": {
                                    "name": "OpenGVLab",
                                    "url": "https://huggingface.co/OpenGVLab"
                                }
                            }
                        ]
                    },
                    "description": "\n\t\n\t\t\n\t\tMMPR-v1.2\n\t\n\n[\ud83d\udcc2 GitHub]  [\ud83c\udd95 Blog]  [\ud83d\udcdc Paper] [\ud83d\udcd6 Documents]\nThis is a newer version of MMPR and MMPR-v1.1, which includes additional data sources to enhance the data diversity and greatly improves the overall performance of InternVL3 across all scales. The prompts used to build this dataset is released in MMPR-v1.2-prompts.\nTo unzip the archive of images, please first run cat images.zip_* > images.zip and then run unzip images.zip.\n\n\n\t\n\t\t\n\t\tIntroduction\n\t\n\nMMPR is a large-scale and\u2026 See the full description on the dataset page: https://huggingface.co/datasets/OpenGVLab/MMPR-v1.2."
                }
            ]
        }
    ]
}