i$zdZddlZddlmZddlmZddlmZddl m Z ddl m Z dd l mZdd lmZmZdd lmZd d lmZmZd dlmZmZmZd dlmZmZmZmZm Z m!Z!m"Z"m#Z#edeGddeZ$edeGddeZ%edeGddeZ&edeGddeZ'GddeZ(Gdde"Z)GddeZ*Gd d!e!Z+Gd"d#eZ,eGd$d%e Z-ed&'Gd(d)e#Z.Gd*d+eZ/gd,Z0y)-zPyTorch SAM 2 model.N)strict)initialization)PreTrainedConfig)PreTrainedModel)Unpack)auto_docstring)TransformersKwargsmerge_with_config_defaults)capture_outputs)CONFIG_MAPPING AutoConfig) Sam2ConfigSam2MaskDecoderConfigSam2PromptEncoderConfig) Sam2AttentionSam2FeedForward Sam2LayerNorm Sam2ModelSam2PreTrainedModelSam2TwoWayAttentionBlockSam2VisionEncoderOutputSam2VisionModelzyonigozlan/EdgeTAM-hf) checkpointceZdZUdZdZdZdeiZdZe e zdze d<dZ e edze d<dZe dze d<dZee d <d Zee d <d Zee d <d Zee d<dZe edze d<dZee d<dZee d<dZee d<dZee d<fdZxZS)EdgeTamVisionConfiga backbone_channel_list (`List[int]`, *optional*, defaults to `[384, 192, 96, 48]`): The list of channel dimensions for the backbone. backbone_feature_sizes (`List[List[int]]`, *optional*, defaults to `[[256, 256], [128, 128], [64, 64]]`): The spatial sizes of the feature maps from the backbone. fpn_hidden_size (`int`, *optional*, defaults to 256): The hidden dimension of the FPN. fpn_kernel_size (`int`, *optional*, defaults to 1): The kernel size for the convolutions in the neck. fpn_stride (`int`, *optional*, defaults to 1): The stride for the convolutions in the neck. fpn_padding (`int`, *optional*, defaults to 0): The padding for the convolutions in the neck. fpn_top_down_levels (`List[int]`, *optional*, defaults to `[2, 3]`): The levels for the top-down FPN connections. num_feature_levels (`int`, *optional*, defaults to 3): The number of feature levels from the FPN to use. vision_configedgetam_vision_modelbackbone_configNbackbone_channel_listbackbone_feature_sizesfpn_hidden_sizefpn_kernel_size fpn_strider fpn_paddingfpn_top_down_levelsrnum_feature_levelsgelu hidden_actgư>layer_norm_epsg{Gz?initializer_rangec 0|jgdn |j|_|j ddgddgddggn |j|_|jddgn |j|_t|jt rT|jj dd|jd<t|jddi|j|_n.|j"tjd dd gd d  |_t|,di|y)N)i`0r#@r r model_type timm_wrapperztimm/repvit_m1.dist_in1kT)rr%r r)in_chans features_only out_indices) model_args) r!r"r) isinstancer dictgetrrfrom_pretrainedsuper __post_init__)selfkwargs __class__s |/var/www/vps2.regionflexible.com/Desarrollo/venv/lib/python3.12/site-packages/transformers/models/edgetam/modular_edgetam.pyrAz!EdgeTamVisionConfig.__post_init__Qs"&"<"<"D $JdJd "372M2M2Uc3Z#sb"X .[_[v[v #.2-E-E-MAq6SWSkSk d**D 1151E1E1I1I,Xf1gD  .#1$2F2F|2T#U#mX\XlXl#mD  ! ! )#-#=#=*()DQ]^$D  '')__name__ __module__ __qualname____doc__base_config_keyr5r sub_configsr r=r__annotations__r!listintr"r$r&r'r(r)r*r,strr-floatr.rA __classcell__)rDs@rErr(s&&O'J:K7;OT,,t3:.249t+2*.D4K.OSOSJK,0cT)0J NE #u#((rFrc eZdZy)EdgeTamPromptEncoderConfigNrGrHrIr;rFrErTrTe rFrTc eZdZy)EdgeTamMaskDecoderConfigNrUr;rFrErXrXkrVrFrXceZdZdZy) EdgeTamConfiga prompt_encoder_config (Union[`dict`, `EdgeTamPromptEncoderConfig`], *optional*): Dictionary of configuration options used to initialize [`EdgeTamPromptEncoderConfig`]. mask_decoder_config (Union[`dict`, `EdgeTamMaskDecoderConfig`], *optional*): Dictionary of configuration options used to initialize [`EdgeTamMaskDecoderConfig`]. Example: ```python >>> from transformers import ( ... EdgeTamVisionConfig, ... EdgeTamPromptEncoderConfig, ... EdgeTamMaskDecoderConfig, ... EdgeTamModel, ... ) >>> # Initializing a EdgeTamConfig with `"facebook/edgetam.1_hiera_tiny"` style configuration >>> configuration = EdgeTamConfig() >>> # Initializing a EdgeTamModel (with random weights) from the `"facebook/edgetam.1_hiera_tiny"` style configuration >>> model = EdgeTamModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config >>> # We can also initialize a EdgeTamConfig from a EdgeTamVisionConfig, EdgeTamPromptEncoderConfig, and EdgeTamMaskDecoderConfig >>> # Initializing EDGETAM vision encoder, memory attention, and memory encoder configurations >>> vision_config = EdgeTamVisionConfig() >>> prompt_encoder_config = EdgeTamPromptEncoderConfig() >>> mask_decoder_config = EdgeTamMaskDecoderConfig() >>> config = EdgeTamConfig(vision_config, prompt_encoder_config, mask_decoder_config) ``` N)rGrHrIrJr;rFrErZrZqs !F rFrZc eZdZy)EdgeTamLayerNormNrUr;rFrEr\r\rFr\c eZdZy)EdgeTamVisionEncoderOutputNrUr;rFrEr_r_r]rFr_c eZdZy)EdgeTamAttentionNrUr;rFrErarar]rFrac eZdZy)EdgeTamTwoWayAttentionBlockNrUr;rFrErcrcr]rFrcc eZdZy)EdgeTamFeedForwardNrUr;rFrErerer]rFrec>eZdZdZej dZy)EdgeTamPreTrainedModelNctj||t|tr-|j t j |jyyt|dr,t j|j|jyy)Npositional_embedding)std) r _init_weightsr< EdgeTamModelno_memory_embeddinginitzeros_hasattrnormal_riscale)rBmodules rErkz$EdgeTamPreTrainedModel._init_weightssg%%dF3 fl +))5 F6676 V3 4 LL44&,, G5rF)rGrHrI"_keys_to_ignore_on_load_unexpectedtorchno_gradrkr;rFrErgrgs$)-&U]]_HHrFrgzN The vision model from EdgeTAM without any head or projection on top. ) custom_introc peZdZeZdZiZdZee dde jdzde e deezfdZy)EdgeTamVisionModel pixel_valuesctdNz2Can't get input embeddings from timm wrapper modelNotImplementedErrorrBs rEget_input_embeddingsz'EdgeTamVisionModel.get_input_embeddings!"VWWrFNrCreturnc ^| td|j|fi|}|j}|Dcgc]}|jdddd}}|j |\}}||j dddd}||j dddd}t |d|||jScc}w)Nz You have to specify pixel_valuesrr rr%)last_hidden_statefpn_hidden_statesfpn_position_encoding hidden_states) ValueErrorbackbonerpermuteneckr*r_r)rBrzrCbackbone_outputintermediate_hidden_states hidden_staterrs rEforwardzEdgeTamVisionModel.forwards  ?@ @($-- ??%4%F%F"[u%vrs~ .&3-&#K5-``   238(*8(48(v23 !8 4 23 4 4 23$ J$ 4$ N }  !8  }  ":    H0 H H # #  # L X9 X  rF