config/eval/msvd_qa/test_config.json

{
    "dataset":{
        "dataset_name": "MSVD_QA",
        "q_json_path":"/home/user/students/ml/dataset/MSVD_Zero_Shot_QA/test_q.json",  
        "a_json_path":"/home/user/students/ml/dataset/MSVD_Zero_Shot_QA/test_a.json",  
        "video_path":"/home/user/students/ml/dataset/MSVD_Zero_Shot_QA/videos",          
        "data_type":"video",                           
        "bound":"False",                                 
        "question_key":"question",                    
        "answer_key":"answer",                         
        "name_key":"video_name",                      
        "video_postfix":[".avi"],
        "num_segments":8
    },
    "model":{
        "model_path":"/home/user/students/ml/model/LLaVA-NeXT-Video-7B-DPO-hf"
    },
    "conv_mode":"llava_next_video_template",
    "save_path":"./result/eval/msvd_qa",
    "experiment_name":"test",
    "description":"simple inference using llava_next_video with llava_next_video_template in msvd",
    "batch_size":4
}