Moonvit SO 400M

🚀 MoonViT

MoonViT is a Native-resolution Vision Encoder, initialized from and continuously pre - trained on SigLIP - SO - 400M, facilitating image feature extraction.

🚀 Quick Start

MoonViT is a Native-resolution Vision Encoder, which is initialized from and continually pre-trained on SigLIP-SO-400M. To facilitate the standalone use of MoonViT, we have separated the implementation and weights of MoonViT from moonshotai/Kimi-VL-A3B-Instruct.

If you are interested in the training process of MoonViT, you are welcome to read Paper Kimi-VL Technical Report.

💻 Usage Examples

Basic Usage

from PIL import Image
from transformers import AutoModel, AutoImageProcessor

model_path = "moonshotai/MoonViT-SO-400M"
model = AutoModel.from_pretrained(
    model_path,
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True,
)
processor = AutoImageProcessor.from_pretrained(model_path, trust_remote_code=True)

image_path = "./figures/demo.png"
image = Image.open(image_path)

images_processed = processor(image, return_tensors="pt").to(dtype=model.dtype, device=model.device)
image_features: list = model(images_processed.pixel_values, images_processed.image_grid_hws)

print(f"dtype: {image_features[0].dtype}, shape: {image_features[0].shape}")
# dtype: torch.bfloat16, shape: torch.Size([1092, 4, 1152])

📄 License

This project is licensed under the MIT license.