# 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 # This file was automatically generated from src/transformers/models/sam3_lite_text/modular_sam3_lite_text.py. # Do NOT edit this file manually as any edits will be overwritten by the generation of # the file from the modular. If any change should be done, please apply the change to the # modular_sam3_lite_text.py file directly. One of our CI enforces this. # 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨 # Copyright 2026 the HuggingFace Team. 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 huggingface_hub.dataclasses import strict from ...configuration_utils import PreTrainedConfig from ...utils import auto_docstring from ..auto import CONFIG_MAPPING, AutoConfig @auto_docstring(checkpoint="yonigozlan/sam3-litetext-s0") @strict class Sam3LiteTextGeometryEncoderConfig(PreTrainedConfig): r""" roi_size (`int`, *optional*, defaults to 7): ROI size for box pooling operations. """ model_type = "sam3_lite_text_geometry_encoder" hidden_size: int = 256 num_layers: int = 3 num_attention_heads: int = 8 intermediate_size: int = 2048 dropout: float | int = 0.1 hidden_act: str = "relu" hidden_dropout: float | int = 0.0 layer_norm_eps: float = 1e-6 roi_size: int = 7 initializer_range: float = 0.02 @auto_docstring(checkpoint="yonigozlan/sam3-litetext-s0") @strict class Sam3LiteTextDETREncoderConfig(PreTrainedConfig): r""" hidden_dropout (`float`, *optional*, defaults to 0.0): Dropout probability for hidden states. """ model_type = "sam3_lite_text_detr_encoder" hidden_size: int = 256 num_layers: int = 6 num_attention_heads: int = 8 intermediate_size: int = 2048 dropout: float | int = 0.1 hidden_act: str = "relu" hidden_dropout: float | int = 0.0 layer_norm_eps: float = 1e-6 initializer_range: float = 0.02 @auto_docstring(checkpoint="yonigozlan/sam3-litetext-s0") @strict class Sam3LiteTextDETRDecoderConfig(PreTrainedConfig): r""" num_queries (`int`, *optional*, defaults to 200): Number of object queries. """ model_type = "sam3_lite_text_detr_decoder" hidden_size: int = 256 num_layers: int = 6 num_queries: int = 200 num_attention_heads: int = 8 intermediate_size: int = 2048 dropout: float | int = 0.1 hidden_act: str = "relu" hidden_dropout: float | int = 0.0 layer_norm_eps: float = 1e-6 initializer_range: float = 0.02 @auto_docstring(checkpoint="yonigozlan/sam3-litetext-s0") @strict class Sam3LiteTextMaskDecoderConfig(PreTrainedConfig): r""" num_upsampling_stages (`int`, *optional*, defaults to 3): Number of upsampling stages in the pixel decoder (FPN). """ model_type = "sam3_lite_text_mask_decoder" hidden_size: int = 256 num_upsampling_stages: int = 3 layer_norm_eps: float = 1e-6 dropout: float | int = 0.0 num_attention_heads: int = 8 initializer_range: float = 0.02 @auto_docstring(checkpoint="yonigozlan/sam3-litetext-s0") @strict class Sam3LiteTextTextConfig(PreTrainedConfig): r""" use_repmixer_blocks (`bool`, *optional*, defaults to `True`): Whether to use RepMixer blocks (MobileCLIP-style) for the first and last encoder layers. When `False`, all layers are standard Transformer encoder layers. layer_scale_init_value (`float`, *optional*, defaults to `1e-5`): Initial value for the learnable layer-scale parameters in RepMixer blocks (residual branches). repmixer_kernel_size (`int`, *optional*, defaults to `11`): Kernel size for depthwise convolutions in RepMixer blocks (token mixer and convolutional feed-forward path). """ model_type = "sam3_lite_text_text_model" vocab_size: int = 49408 hidden_size: int = 512 intermediate_size: int = 2048 projection_dim: int = 512 num_hidden_layers: int = 12 num_attention_heads: int = 8 max_position_embeddings: int = 77 hidden_act: str = "gelu" layer_norm_eps: float = 1e-5 attention_dropout: float = 0.0 use_repmixer_blocks: bool = True layer_scale_init_value: float = 1e-5 repmixer_kernel_size: int = 11 @auto_docstring(checkpoint="yonigozlan/sam3-litetext-s0") @strict class Sam3LiteTextConfig(PreTrainedConfig): r""" geometry_encoder_config (`dict` or `Sam3LiteTextGeometryEncoderConfig`, *optional*): Configuration for the geometry encoder. detr_encoder_config (`dict` or `Sam3LiteTextDETREncoderConfig`, *optional*): Configuration for the DETR encoder. detr_decoder_config (`dict` or `Sam3LiteTextDETRDecoderConfig`, *optional*): Configuration for the DETR decoder. mask_decoder_config (`dict` or `Sam3LiteTextMaskDecoderConfig`, *optional*): Configuration for the mask decoder. Example: ```python >>> from transformers import Sam3LiteTextConfig, Sam3LiteTextModel >>> # Initializing a SAM3_LITE_TEXT configuration >>> configuration = Sam3LiteTextConfig() >>> # Initializing a model from the configuration >>> model = Sam3LiteTextModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` """ model_type = "sam3_lite_text" sub_configs = { "vision_config": AutoConfig, "text_config": Sam3LiteTextTextConfig, "geometry_encoder_config": Sam3LiteTextGeometryEncoderConfig, "detr_encoder_config": Sam3LiteTextDETREncoderConfig, "detr_decoder_config": Sam3LiteTextDETRDecoderConfig, "mask_decoder_config": Sam3LiteTextMaskDecoderConfig, } vision_config: dict | PreTrainedConfig | None = None text_config: dict | PreTrainedConfig | None = None geometry_encoder_config: dict | PreTrainedConfig | None = None detr_encoder_config: dict | PreTrainedConfig | None = None detr_decoder_config: dict | PreTrainedConfig | None = None mask_decoder_config: dict | PreTrainedConfig | None = None initializer_range: float = 0.02 def __post_init__(self, **kwargs): if isinstance(self.vision_config, dict): self.vision_config["model_type"] = self.vision_config.get("model_type", "sam3_vision_model") self.vision_config = CONFIG_MAPPING[self.vision_config["model_type"]](**self.vision_config) elif self.vision_config is None: self.vision_config = CONFIG_MAPPING["sam3_vision_model"]() if self.text_config is None: self.text_config = Sam3LiteTextTextConfig() if isinstance(self.text_config, dict): self.text_config = Sam3LiteTextTextConfig(**self.text_config) if self.geometry_encoder_config is None: self.geometry_encoder_config = Sam3LiteTextGeometryEncoderConfig() if isinstance(self.geometry_encoder_config, dict): self.geometry_encoder_config = Sam3LiteTextGeometryEncoderConfig(**self.geometry_encoder_config) if self.detr_encoder_config is None: self.detr_encoder_config = Sam3LiteTextDETREncoderConfig() if isinstance(self.detr_encoder_config, dict): self.detr_encoder_config = Sam3LiteTextDETREncoderConfig(**self.detr_encoder_config) if self.detr_decoder_config is None: self.detr_decoder_config = Sam3LiteTextDETRDecoderConfig() if isinstance(self.detr_decoder_config, dict): self.detr_decoder_config = Sam3LiteTextDETRDecoderConfig(**self.detr_decoder_config) if self.mask_decoder_config is None: self.mask_decoder_config = Sam3LiteTextMaskDecoderConfig() if isinstance(self.mask_decoder_config, dict): self.mask_decoder_config = Sam3LiteTextMaskDecoderConfig(**self.mask_decoder_config) super().__post_init__(**kwargs) @property def image_size(self): """Image size for the SAM3_LITE_TEXT model.""" return self.vision_config.image_size @image_size.setter def image_size(self, value): """Set the image size and propagate to vision config.""" self.vision_config.image_size = value __all__ = [ "Sam3LiteTextConfig", "Sam3LiteTextTextConfig", "Sam3LiteTextGeometryEncoderConfig", "Sam3LiteTextDETREncoderConfig", "Sam3LiteTextDETRDecoderConfig", "Sam3LiteTextMaskDecoderConfig", ]