--- license: apache-2.0 datasets: - lmms-lab/LLaVA-OneVision-Data language: - en - zh metrics: - accuracy library_name: transformers tags: - multimodal model-index: - name: llava-onevision-qwen-7b-ov results: - task: type: multimodal dataset: type: ai2d name: AI2D metrics: - name: accuracy type: accuracy value: 81.4 verified: true - task: type: multimodal dataset: type: chartqa name: ChartQA metrics: - name: accuracy type: accuracy value: 80.0 verified: true - task: type: multimodal dataset: type: docvqa name: DocVQA metrics: - name: accuracy type: accuracy value: 90.2 verified: true - task: type: multimodal dataset: type: infovqa name: InfoVQA metrics: - name: accuracy type: accuracy value: 70.7 verified: true - task: type: multimodal dataset: type: mathverse name: MathVerse metrics: - name: accuracy type: accuracy value: 26.2 verified: true - task: type: multimodal dataset: type: mathvista name: MathVista metrics: - name: accuracy type: accuracy value: 63.2 verified: true - task: type: multimodal dataset: type: mmbench name: MMBench metrics: - name: accuracy type: accuracy value: 80.8 verified: true - task: type: multimodal dataset: type: mme-perception name: MME-Perception metrics: - name: score type: score value: 1580 verified: true - task: type: multimodal dataset: type: mme-cognition name: MME-Cognition metrics: - name: score type: score value: 418 verified: true - task: type: multimodal dataset: type: mmmu name: MMMU metrics: - name: accuracy type: accuracy value: 48.8 verified: true - task: type: multimodal dataset: type: mmvet name: MMVet metrics: - name: accuracy type: accuracy value: 57.5 verified: true - task: type: multimodal dataset: type: mmstar name: MMStar metrics: - name: accuracy type: accuracy value: 61.7 verified: true - task: type: multimodal dataset: type: seed-bench name: Seed-Bench metrics: - name: accuracy type: accuracy value: 75.4 verified: true - task: type: multimodal dataset: type: science-qa name: Science-QA metrics: - name: accuracy type: accuracy value: 96.0 verified: true - task: type: multimodal dataset: type: imagedc name: ImageDC metrics: - name: accuracy type: accuracy value: 88.9 verified: true - task: type: multimodal dataset: type: mmlbench name: MMLBench metrics: - name: accuracy type: accuracy value: 77.1 verified: true - task: type: multimodal dataset: type: realworldqa name: RealWorldQA metrics: - name: accuracy type: accuracy value: 66.3 verified: true - task: type: multimodal dataset: type: vibe-eval name: Vibe-Eval metrics: - name: accuracy type: accuracy value: 51.7 verified: true - task: type: multimodal dataset: type: llava-w name: LLaVA-W metrics: - name: accuracy type: accuracy value: 90.7 verified: true - task: type: multimodal dataset: type: l-wilder name: LLaVA-Wilder metrics: - name: accuracy type: accuracy value: 67.8 verified: true - task: type: multimodal dataset: type: actnet-qa name: ActNet-QA metrics: - name: accuracy type: accuracy value: 56.6 verified: true - task: type: multimodal dataset: type: egoschema name: EgoSchema metrics: - name: accuracy type: accuracy value: 60.1 verified: true - task: type: multimodal dataset: type: mlvu name: MLVU metrics: - name: accuracy type: accuracy value: 64.7 verified: true - task: type: multimodal dataset: type: mvbench name: MVBench metrics: - name: accuracy type: accuracy value: 56.7 verified: true - task: type: multimodal dataset: type: nextqa name: NextQA metrics: - name: accuracy type: accuracy value: 79.4 verified: true - task: type: multimodal dataset: type: percepTest name: PercepTest metrics: - name: accuracy type: accuracy value: 49.7 verified: true - task: type: multimodal dataset: type: seedbench name: SeedBench metrics: - name: accuracy type: accuracy value: 56.9 verified: true - task: type: multimodal dataset: type: videochatgpt name: VideoChatGPT metrics: - name: score type: score value: 3.49 verified: true - task: type: multimodal dataset: type: videodc name: VideoDC metrics: - name: score type: score value: 3.75 verified: true - task: type: multimodal dataset: type: videomme name: VideoMME metrics: - name: accuracy type: accuracy value: 58.2 verified: true --- # LLaVA-OneVision ![banner](https://i.postimg.cc/pL17YtG4/WX20240508-220230-2x.png) Play with the model on the [LLaVA OneVision Chat](https://llava-onevision.lmms-lab.com/). ## Table of Contents 1. [Model Summary](##model-summary) 2. [Use](##use) 3. [Limitations](##limitations) 4. [Training](##training) 5. [License](##license) 6. [Citation](##citation) ## Model Summary The LLaVA-OneVision models are 0.5/7/72B parameter models trained on [LLaVA-OneVision](https://huggingface.co/datasets/lmms-lab/LLaVA-OneVision-Data), based on Qwen2 language model with a context window of 32K tokens. - **Repository:** [LLaVA-VL/LLaVA-NeXT](https://github.com/LLaVA-VL/LLaVA-NeXT?tab=readme-ov-file) - **Project Website:** [llava-onevision.lmms-lab.com](llava-onevision.lmms-lab.com) - **Paper:** [LLaVA-OneVision]() - **Point of Contact:** [Bo Li](mailto:drluodian@gmail.com) - **Languages:** English, Chinese ## Use ### Intended use The model was trained on [LLaVA-OneVision Dataset](https://huggingface.co/datasets/lmms-lab/LLaVA-OneVision-Data) and have the ability to interact with images, multi-image and videos. **Feel free to share your generations in the Community tab!** ### Generation ```python # pip install git+https://github.com/LLaVA-VL/LLaVA-NeXT.git from llava.model.builder import load_pretrained_model from llava.mm_utils import get_model_name_from_path, process_images, tokenizer_image_token from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN, IGNORE_INDEX from llava.conversation import conv_templates, SeparatorStyle from PIL import Image import requests import copy import torch import sys import warnings warnings.filterwarnings("ignore") pretrained = "lmms-lab/llava-onevision-qwen2-7b-ov" model_name = "llava_qwen" device = "cuda" device_map = "auto" tokenizer, model, image_processor, max_length = load_pretrained_model(pretrained, None, model_name, device_map=device_map) # Add any other thing you want to pass in llava_model_args model.eval() url = "https://github.com/haotian-liu/LLaVA/blob/1a91fc274d7c35a9b50b3cb29c4247ae5837ce39/images/llava_v1_5_radar.jpg?raw=true" image = Image.open(requests.get(url, stream=True).raw) image_tensor = process_images([image], image_processor, model.config) image_tensor = [_image.to(dtype=torch.float16, device=device) for _image in image_tensor] conv_template = "qwen_1_5" # Make sure you use correct chat template for different models question = DEFAULT_IMAGE_TOKEN + "\nWhat is shown in this image?" conv = copy.deepcopy(conv_templates[conv_template]) conv.append_message(conv.roles[0], question) conv.append_message(conv.roles[1], None) prompt_question = conv.get_prompt() input_ids = tokenizer_image_token(prompt_question, tokenizer, IMAGE_TOKEN_INDEX, return_tensors="pt").unsqueeze(0).to(device) image_sizes = [image.size] cont = model.generate( input_ids, images=image_tensor, image_sizes=image_sizes, do_sample=False, temperature=0, max_new_tokens=4096, ) text_outputs = tokenizer.batch_decode(cont, skip_special_tokens=True) print(text_outputs) ``` # Training ## Model - **Architecture:** SO400M + Qwen2 - **Pretraining Stage:** LCS-558K, 1 epoch, projector - **Mid Stage:** A mixture of 4.7M high-quality synthetic data, 1 epoch, full model - **Final-Image Stage:** A mixture of 3.6M single-image data, 1 epoch, full model - **OneVision Stage:** A mixture of 1.6M single-image/multi-image/video data, 1 epoch, full model - **Precision:** bfloat16 ## Hardware & Software - **GPUs:** 256 * Nvidia Tesla A100 (for whole model series training) - **Orchestration:** [Huggingface Trainer](https://huggingface.co/docs/transformers/main_classes/trainer) - **Neural networks:** [PyTorch](https://github.com/pytorch/pytorch) # Citation ``` @article{li2024llavaonevision, title={LLaVA-OneVision}, } ```