configuration_tapct.py

# Copyright 2025 AI for Oncology Research Group. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import Literal
from transformers import PretrainedConfig

class TAPCTConfig(PretrainedConfig):
    """
    Configuration class for TAP-CT models.
    
    Parameters
    ----------
    model_size : Literal['small', 'base'], default='base'
        Size of the model ('small' or 'base')
    model_variant : Literal['2d', '2.5d', '3d'], default='3d'
        Variant of the model ('2d', '2.5d', or '3d')
    img_size : int | tuple | list, default=224
        Input image size. For 2D: int or tuple[int, int], for 3D: tuple[int, int, int]
    patch_size : int | tuple | list, default=16
        Patch size. For 2D: int or tuple[int, int], for 3D: tuple[int, int, int]
    in_chans : int, default=1
        Number of input channels (default: 1 for CT scans)
    num_register_tokens : int, default=4
        Number of register tokens
    init_values : float | None, default=None
        Layer scale init values
    block_chunks : int, default=0
        Number of block chunks for FSDP
    """
    model_type = "tapct"

    def __init__(
        self,
        model_size: Literal['small', 'base'] = 'base',
        model_variant: Literal['2d', '2.5d', '3d'] = '3d',
        img_size: int | tuple | list = 224,
        patch_size: int | tuple | list = 16,
        in_chans: int = 1,
        num_register_tokens: int = 4,
        init_values: float | None = None,
        block_chunks: int = 0,
        **kwargs
    ):
        super().__init__(**kwargs)
        self.model_size = model_size
        self.model_variant = model_variant
        self.img_size = img_size
        self.patch_size = patch_size
        self.in_chans = in_chans
        self.num_register_tokens = num_register_tokens
        self.init_values = init_values
        self.block_chunks = block_chunks