- [2025-09-27] Open-source evaluation codes.
- [2025-09-26] Upload to Arxiv.
# Clone repository
git clone [email protected]:CHIzhP/Chimera.git
cd Chimera
# Create environment
conda create -n chimera python=3.10 -y
conda activate chimera
# Install dependencies
pip install -r requirements.txtcd evaluation
python run.py --lvlm_path <your_model_path> [other arguments...]-
--lvlm_path: Specify the path of the model to be evaluated.
-
--API_KEY: Provide an API key if the model requires API-based access.
-
--modality: Choose the input modality. (Options:
visual|semantic|textual|empty) -
--task: Select the evaluation task type. (Options:
Recognition|Understanding|Grounding|Reasoning|all)