config.json

{
  "architectures": [
    "PPOCRV6MediumDetForObjectDetection"
  ],
  "backbone_config": {
    "block_configs": [
      [
        [
          3,
          16,
          16,
          1,
          false
        ]
      ],
      [
        [
          3,
          16,
          16,
          2,
          false
        ],
        [
          3,
          16,
          16,
          1,
          false
        ]
      ],
      [
        [
          3,
          16,
          16,
          2,
          false
        ],
        [
          3,
          16,
          16,
          1,
          false
        ]
      ],
      [
        [
          3,
          16,
          16,
          2,
          false
        ],
        [
          5,
          16,
          16,
          1,
          false
        ],
        [
          5,
          16,
          16,
          1,
          false
        ],
        [
          5,
          16,
          16,
          1,
          false
        ],
        [
          5,
          16,
          16,
          1,
          false
        ]
      ]
    ],
    "depths": [
      1,
      2,
      2,
      5
    ],
    "hidden_act": "relu",
    "model_type": "pp_lcnet_v4",
    "num_channels": 3,
    "out_features": [
      "stage1",
      "stage2",
      "stage3",
      "stage4"
    ],
    "out_indices": [
      1,
      2,
      3,
      4
    ],
    "reduction": 4,
    "scale": 1.0,
    "stage_names": [
      "stem",
      "stage1",
      "stage2",
      "stage3",
      "stage4"
    ],
    "stage_out_channels": [
      16,
      16,
      16,
      16
    ],
    "stem_channels": [
      3,
      16,
      16
    ],
    "stem_strides": [
      2,
      1,
      1,
      2,
      1
    ],
    "stem_type": "large",
    "use_learnable_affine_block": false
  },
  "dtype": "float32",
  "hidden_act": "relu",
  "id2label": {
    "0": "text"
  },
  "interpolate_mode": "nearest",
  "intraclass_block_config": {
    "horizontal_small_to_long_conv_longratio": [
      [
        1,
        7
      ],
      [
        1,
        1
      ],
      [
        0,
        3
      ]
    ],
    "horizontal_small_to_long_conv_midratio": [
      [
        1,
        5
      ],
      [
        1,
        1
      ],
      [
        0,
        2
      ]
    ],
    "horizontal_small_to_long_conv_shortratio": [
      [
        1,
        3
      ],
      [
        1,
        1
      ],
      [
        0,
        1
      ]
    ],
    "reduce_channel": [
      1,
      1,
      0
    ],
    "return_channel": [
      1,
      1,
      0
    ],
    "symmetric_conv_long_longratio": [
      [
        7,
        7
      ],
      [
        1,
        1
      ],
      [
        3,
        3
      ]
    ],
    "symmetric_conv_long_midratio": [
      [
        5,
        5
      ],
      [
        1,
        1
      ],
      [
        2,
        2
      ]
    ],
    "symmetric_conv_long_shortratio": [
      [
        3,
        3
      ],
      [
        1,
        1
      ],
      [
        1,
        1
      ]
    ],
    "vertical_long_to_small_conv_longratio": [
      [
        7,
        1
      ],
      [
        1,
        1
      ],
      [
        3,
        0
      ]
    ],
    "vertical_long_to_small_conv_midratio": [
      [
        5,
        1
      ],
      [
        1,
        1
      ],
      [
        2,
        0
      ]
    ],
    "vertical_long_to_small_conv_shortratio": [
      [
        3,
        1
      ],
      [
        1,
        1
      ],
      [
        1,
        0
      ]
    ]
  },
  "intraclass_block_number": 4,
  "kernel_list": [
    3,
    2,
    2
  ],
  "label2id": null,
  "mode": "large",
  "model_type": "pp_ocrv6_medium_det",
  "neck_out_channels": 16,
  "reduce_factor": 2,
  "scale_factor": 2,
  "scale_factor_list": [
    1,
    2,
    4,
    8
  ],
  "transformers_version": "5.16.0.dev0"
}