Sam2 Hiera Large

🚀 SAM2-Hiera-large

This repository houses the large variant of the SAM2 model. SAM2 is the state-of-the-art mask generation model released by Meta, offering high - performance solutions for mask generation tasks.

🚀 Quick Start

To start using the SAM2-Hiera-large model, you first need to install the packaged version of SAM2.

📦 Installation

pip install samv2 huggingface_hub

💻 Usage Examples

Basic Usage

Each model requires different classes for inference. The following code demonstrates how to use the SAM2ImagePredictor for mask prediction.

from huggingface_hub import hf_hub_download
from sam2.build_sam import build_sam2
from sam2.sam2_image_predictor import SAM2ImagePredictor

hf_hub_download(repo_id = "merve/sam2-hiera-large", filename="sam2_hiera_large.pt", local_dir = "./")
sam2_checkpoint = "../checkpoints/sam2_hiera_large.pt"
model_cfg = "sam2_hiera_l.yaml"

sam2_model = build_sam2(config, ckpt, device="cuda", apply_postprocessing=False)
predictor = SAM2ImagePredictor(sam2_model)

# it accepts coco format
box = [x1, y1, w, h]
predictor.set_image(image)

masks = predictor.predict(box=box,
multimask_output=False)

Advanced Usage

For automatic mask generation, you can use the SAM2AutomaticMaskGenerator class as shown below.

from huggingface_hub import hf_hub_download
from sam2.build_sam import build_sam2
from sam2.automatic_mask_generator import SAM2AutomaticMaskGenerator

hf_hub_download(repo_id = "merve/sam2-hiera-large", filename="sam2_hiera_large.pt", local_dir = "./")
sam2_checkpoint = "../checkpoints/sam2_hiera_large.pt"
model_cfg = "sam2_hiera_l.yaml"

sam2 = build_sam2(model_cfg, sam2_checkpoint, device ='cuda', apply_postprocessing=False)

mask_generator = SAM2AutomaticMaskGenerator(sam2)
masks = mask_generator.generate(image)

📚 Documentation

The team behind SAM2 has created example notebooks for all tasks. You can refer to the following links for more detailed examples:

• Image predictor example: This notebook provides a full example on prompting.
• Automatic mask generation example: Use this notebook to generate all masks.
• Video object segmentation example

📄 License

This project is licensed under the Apache-2.0 license.