Fi<2 $ddlmZddlZddlmZddlmZddlmZddl m Z m Z m Z m Z mZmZddlmZmZdd lmZmZdd lmZdd lmZd Zdd ZddZddZddZddZddZ ddZ!ddZ" ddZ# ddZ$eeeeee e!e"ee e!e"d Z%ddZ&y))partialN)attempt_download_asset) torch_load) MaskDecoder)FpnNeckHiera ImageEncoderImageEncoderViT MemoryEncoder PromptEncoder)MemoryAttentionMemoryAttentionLayer) SAM2ModelSAMModel)TinyViT)TwoWayTransformerc||St|}t|d5}t|}ddddvrt|dtr|d}|j ||S#1swY8xYw)z*Load checkpoint into model from file path.Nrbmodel)ropenr isinstancedictload_state_dict)r checkpointf state_dicts ^/var/www/pod-logistic/pod-ai/venv/lib/python3.12/site-packages/ultralytics/models/sam/build.py_load_checkpointrsv ' 3J j$ #1] #*Jw,?!F(  *% L ##s AA(c&tdddgd|S)z_Build and return a Segment Anything Model (SAM) h-size model with specified encoder parameters. encoder_embed_dim encoder_depthencoder_num_headsencoder_global_attn_indexesr _build_samrs rbuild_sam_vit_hr1' $3  c&tdddgd|S)z_Build and return a Segment Anything Model (SAM) l-size model with specified encoder parameters.r#) r'r)r.r0s rbuild_sam_vit_lr:2r2r3c&tdddgd|S)z_Build and return a Segment Anything Model (SAM) b-size model with specified encoder parameters. )r7r8r)r.r0s rbuild_sam_vit_br@=s $1  r3c0tgdgdgddd|S)z_Build and return a Mobile Segment Anything Model (Mobile-SAM) for efficient image segmentation.)@i@)r>r>r>)r>r7 NT)r*r+r,r- mobile_samrr.r0s rbuild_mobile_samrIHs! -"'$(  r3c 6tdgddgdgdgd|S)zjBuild and return a Segment Anything Model 2 (SAM2) tiny-size model with specified architecture parameters.`)rr>r%r>r)r7r% r?rFr%r<irKr*encoder_stagesr,encoder_global_att_blocksencoder_window_specencoder_backbone_channel_listr _build_sam2r0s r build_sam2_trXTs$ #"+)&9 r3c 6tdgddgdgdgd|S)zeBuild and return a small-size Segment Anything Model 2 (SAM2) with specified architecture parameters.rK)rr>r8r>r)r%rG rMrOrQrVr0s r build_sam2_sr[as$ $"-)&9 r3c <tdgddgdgdddggd|S) zjBuild and return a Segment Anything Model 2 (SAM2) base-size model with specified architecture parameters.p)r>r#r^r>)r=r#rMrN)iir])r*rRr,rSrTencoder_window_spatial_sizerUrrVr0s r build_sam2_brbns+ $".)%'H&:  r3c 6tdgddgdgdgd|S)zeBuild and return a large-size Segment Anything Model 2 (SAM2) with specified architecture parameters.r>rE$rFr>)r'!+r?rFr#r?ii@i rdrQrVr0s r build_sam2_lrk|s$ $".)&; r3cd}d}d}||z} |rtddd|||gddddd ddd n9t|||d ttjj d ||dd|d| } t | t|| | f||fdtdtd|dd|ddgdgd} | t| |} | j| S)aOBuild a Segment Anything Model (SAM) with specified encoder parameters. Args: encoder_embed_dim (int | list[int]): Embedding dimension for the encoder. encoder_depth (int | list[int]): Depth of the encoder. encoder_num_heads (int | list[int]): Number of attention heads in the encoder. encoder_global_attn_indexes (list[int] | None): Indexes for global attention in the encoder. checkpoint (str | None, optional): Path to the model checkpoint file. mobile_sam (bool, optional): Whether to build a Mobile-SAM model. Returns: (SAMModel): A Segment Anything Model instance with the specified architecture. Examples: >>> sam = _build_sam(768, 12, 12, [2, 5, 8, 11]) >>> sam = _build_sam([64, 128, 160, 320], [2, 2, 6, 2], [2, 4, 5, 10], None, mobile_sam=True) r5r#r^i)r%r%rNr%g@gFg?)img_sizein_chans num_classes embed_dimsdepths num_heads window_sizes mlp_ratio drop_ratedrop_path_rateuse_checkpointmbconv_expand_ratiolocal_conv_sizelayer_lr_decayrFgư>)epsTrN) depth embed_dimrnru norm_layerrs patch_sizeqkv_bias use_rel_posglobal_attn_indexes window_size out_chans)r~image_embedding_sizeinput_image_size mask_in_chansr>ir?)r} embedding_dimmlp_dimrs)num_multimask_outputs transformertransformer_dimiou_head_depthiou_head_hidden_dim)g33333^@gR]@gRY@)g(\2M@g(\L@gL@) image_encoderprompt_encoder mask_decoder pixel_mean pixel_std) rr rtorchnn LayerNormrr rrreval) r*r+r,r-rrHprompt_embed_dim image_sizevit_patch_sizerrsams rr/r/s!2JN%7$ ! ( '& # "'uxx11t<'% ;&  %B #$&"68L!M(*5  !"#).  - #  -)+ C.sJ/HHJ Jr3c tt||||||td|ddgdd}tdd d t } t d } |duxrd|v} t d5id|d| d| dddddddddd dd dd dd dd d d d!d d"d d#d d$d d%d d&d d'd d(d)d*dd+d d,d-d.| d/| d0| d1td d2d34} | t| |} | j| S)6a`Build and return a Segment Anything Model 2 (SAM2) with specified architecture parameters. Args: encoder_embed_dim (int, optional): Embedding dimension for the encoder. encoder_stages (list[int], optional): Number of blocks in each stage of the encoder. encoder_num_heads (int, optional): Number of attention heads in the encoder. encoder_global_att_blocks (list[int], optional): Indices of global attention blocks in the encoder. encoder_backbone_channel_list (list[int], optional): Channel dimensions for each level of the encoder backbone. encoder_window_spatial_size (list[int], optional): Spatial size of the window for position embeddings. encoder_window_spec (list[int], optional): Window specifications for each stage of the encoder. checkpoint (str | None, optional): Path to the checkpoint file for loading pre-trained weights. Returns: (SAM2Model): A configured and initialized SAM2 model. Examples: >>> sam2_model = _build_sam2(encoder_embed_dim=96, encoder_stages=[1, 2, 7, 2]) >>> sam2_model.eval() )r~rsstagesglobal_att_blocks!window_pos_embed_bkg_spatial_size window_specrmr>r^nearest)d_modelbackbone_channel_listfpn_top_down_levelsfpn_interp_modelr)trunkneckscalpTrF)rpos_enc_at_input num_layerslayerrB)out_dimNzsam2.1rmemory_attentionmemory_encoder num_maskmemr%rr5sigmoid_scale_for_mem_encg4@sigmoid_bias_for_mem_encg$$use_mask_input_as_output_without_samdirectly_add_no_mem_embeduse_high_res_features_in_sammultimask_output_in_samiou_prediction_use_sigmoiduse_obj_ptrs_in_encoderadd_tpos_enc_to_obj_ptrs"only_obj_ptrs_in_the_past_for_evalpred_obj_scorespred_obj_scores_mlpfixed_no_obj_ptrmultimask_output_for_trackinguse_multimask_token_for_obj_ptrmultimask_min_pt_numrmultimask_max_pt_numuse_mlp_for_obj_ptr_projcompile_image_encoderFno_obj_embed_spatialproj_tpos_enc_in_obj_ptrsuse_signed_tpos_enc_to_obj_ptrssam_mask_decoder_extra_argsg?g\(\?)dynamic_multimask_via_stability!dynamic_multimask_stability_delta"dynamic_multimask_stability_thresh) r r rrrr rrrr) r*rRr,rSrUrarTrrrr is_sam2_1sam2s rrWrWs:!''!7.I+  "?!"A&  M"'sTVW_s_uv"2.N$&A8z+AI ! #! )! &!  !   !  #' ! "'! .2! #'! &*! !%! $(! !%! "&! ,0!  !! "!#! $%! &'+'! ()-)! *+! ,-! ."&/! 0$1! 2'3! 4#,5! 6)27! 8%),0.2/3% 9! DFj1IIK Kr3) zsam_h.ptzsam_l.ptsam_b.ptz mobile_sam.ptz sam2_t.ptz sam2_s.ptz sam2_b.ptz sam2_l.ptz sam2.1_t.ptz sam2.1_s.ptz sam2.1_b.ptz sam2.1_l.ptcd}t|}tD])}|j|stj|}+|s"t |dtj ||S)aBuild and return a Segment Anything Model (SAM) based on the provided checkpoint. Args: ckpt (str | Path, optional): Path to the checkpoint file or name of a pre-defined SAM model. Returns: (SAMModel | SAM2Model): A configured and initialized SAM or SAM2 model instance. Raises: FileNotFoundError: If the provided checkpoint is not a supported SAM model. Examples: >>> sam_model = build_sam("sam_b.pt") >>> sam_model = build_sam("path/to/custom_checkpoint.pt") Notes: Supported pre-defined models include: - SAM: 'sam_h.pt', 'sam_l.pt', 'sam_b.pt', 'mobile_sam.pt' - SAM2: 'sam2_t.pt', 'sam2_s.pt', 'sam2_b.pt', 'sam2_l.pt' Nz7 is not a supported SAM model. Available models are: )str sam_model_mapendswithgetFileNotFoundErrorkeys)ckpt model_builderks r build_samrOso*M t9D 1 == )--a0M1 4&(`anasasau`v wxx  r3)N)NF)r!rer>r$rj)r%r%riN)r)' functoolsrrultralytics.utils.downloadsrultralytics.utils.patchesrmodules.decodersrmodules.encodersrr r r r r modules.memory_attentionrr modules.samrrmodules.tiny_encoderrmodules.transformerrrr1r:r@rIrXr[rbrkr/rWrrrr3rrs >0)iiK,)2      $ Xx -"7 &%Xx %  r3