Model Card for InternViT-6B-224px

[InternVL 1.5 Technical Report] [Paper] [GitHub] [Chat Demo] [中文解读]

Model Details

• Model Type: vision foundation model, feature backbone
• Model Stats:
  • Params (M): 5903
  • Image size: 224 x 224
• Pretrain Dataset: LAION-en, LAION-COCO, COYO, CC12M, CC3M, SBU, Wukong, LAION-multi
• Note: This model has 48 blocks, and we found that using the output after the fourth-to-last block worked best for VLLM. Therefore, when building a VLLM with this model, please use the features from the fourth-to-last layer.

Linear Probing Performance

See this document for more details about the linear probing evaluation.

Model Usage (Image Embeddings)

import torch
from PIL import Image
from transformers import AutoModel, CLIPImageProcessor

model = AutoModel.from_pretrained(
    'OpenGVLab/InternViT-6B-224px',
    torch_dtype=torch.bfloat16,
    low_cpu_mem_usage=True,
    trust_remote_code=True).cuda().eval()

image = Image.open('./examples/image1.jpg').convert('RGB')

image_processor = CLIPImageProcessor.from_pretrained('OpenGVLab/InternViT-6B-224px')

pixel_values = image_processor(images=image, return_tensors='pt').pixel_values
pixel_values = pixel_values.to(torch.bfloat16).cuda()

outputs = model(pixel_values)

Citation

If you find this project useful in your research, please consider citing:

@article{chen2023internvl,
  title={InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks},
  author={Chen, Zhe and Wu, Jiannan and Wang, Wenhai and Su, Weijie and Chen, Guo and Xing, Sen and Zhong, Muyan and Zhang, Qinglong and Zhu, Xizhou and Lu, Lewei and Li, Bin and Luo, Ping and Lu, Tong and Qiao, Yu and Dai, Jifeng},
  journal={arXiv preprint arXiv:2312.14238},
  year={2023}
}

Acknowledgement

InternVL is built with reference to the code of the following projects: OpenAI CLIP, Open CLIP, CLIP Benchmark, EVA, InternImage, ViT-Adapter, MMSegmentation, Transformers, DINOv2, BLIP-2, Qwen-VL, and LLaVA-1.5. Thanks for their awesome work!