config.json

{
  "_device": "cuda",
  "alpha": 0.5,
  "architectures": [
    "VineModel"
  ],
  "auto_map": {
    "AutoConfig": "vine_config.VineConfig",
    "AutoModel": "vine_model.VineModel"
  },
  "bbox_min_dim": 5,
  "box_threshold": 0.35,
  "custom_pipelines": {
    "vine-video-understanding": {
      "impl": "vine_pipeline.VinePipeline",
      "pt": [
        "VineModel"
      ],
      "tf": [],
      "type": "multimodal"
    }
  },
  "debug_visualizations": false,
  "hidden_dim": 768,
  "interested_object_pairs": [],
  "max_video_length": 100,
  "model_name": "openai/clip-vit-base-patch32",
  "model_type": "vine",
  "multi_class": false,
  "num_top_pairs": 18,
  "output_logit": false,
  "use_hf_repo": false,
  "model_repo": "KevinX-Penn28/testing",
  "model_file": null,
  "local_dir": "./",
  "local_filename": "laser_model_v1.pkl",
  "return_flattened_segments": false,
  "return_valid_pairs": false,
  "segmentation_method": "grounding_dino_sam2",
  "target_fps": 1,
  "text_threshold": 0.25,
  "topk_cate": 3,
  "torch_dtype": "float32",
  "transformers_version": "4.46.2",
  "visualization_dir": null,
  "visualize": false,
  "white_alpha": 0.8
}