AutoMRM is a user-task-centric multimodal model retrieval method . Here is the implementation of the paper "AutoMRM: A Model Retrieval Method Based on Multimodal Query and Metalearning".
It consists of the following parts:
- Metafeature: The feature extraction of the proposed method and baselines.
- Evaluation: The performance evaluation of all candidate models on different tasks.
- Training&Test: The training and test of meta models.
##Requirements: Different parts of the code depend on different environments, please refer to readme and import: python = 3.7.13 pytorch = 1.13.0 clip = 1.0 numpy = 1.21.6 scikit-learn = 1.0.2