wi@ddlmZddlmZmZmZddlZddlmZddl m Z m Z ddl m Z ddlmZmZmZdd lmZd d lmZmZmZeGd d e ZGddee ZdS)) dataclass)OptionalTupleUnionN) ConfigMixinregister_to_config) BaseOutput)GaussianFourierProjectionTimestepEmbedding Timesteps) ModelMixin)UNetMidBlock2Dget_down_block get_up_blockc(eZdZUdZejed<dS) UNet2DOutputz The output of [`UNet2DModel`]. Args: sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)`): The hidden states output from the last layer of the model. sampleN)__name__ __module__ __qualname____doc__torchTensor__annotations__n/root/.openclaw/workspace/chatterbox_venv_py311/lib/python3.11/site-packages/diffusers/models/unets/unet_2d.pyrrs, Lrrc7eZdZdZe d6deeeeeeffdedede de dede dee dfdee dfdeedfded e d!ed"e d#e d$e d%e d&eed'ed(eed)e d*e d+e d,ee d-eed.eef4fd/ Z d7d0e jd1ee je efd2ee jd3e d4eeeff d5ZxZS)8 UNet2DModela A 2D UNet model that takes a noisy sample and a timestep and returns a sample shaped output. This model inherits from [`ModelMixin`]. Check the superclass documentation for it's generic methods implemented for all models (such as downloading or saving). Parameters: sample_size (`int` or `Tuple[int, int]`, *optional*, defaults to `None`): Height and width of input/output sample. Dimensions must be a multiple of `2 ** (len(block_out_channels) - 1)`. in_channels (`int`, *optional*, defaults to 3): Number of channels in the input sample. out_channels (`int`, *optional*, defaults to 3): Number of channels in the output. center_input_sample (`bool`, *optional*, defaults to `False`): Whether to center the input sample. time_embedding_type (`str`, *optional*, defaults to `"positional"`): Type of time embedding to use. freq_shift (`int`, *optional*, defaults to 0): Frequency shift for Fourier time embedding. flip_sin_to_cos (`bool`, *optional*, defaults to `True`): Whether to flip sin to cos for Fourier time embedding. down_block_types (`Tuple[str]`, *optional*, defaults to `("DownBlock2D", "AttnDownBlock2D", "AttnDownBlock2D", "AttnDownBlock2D")`): Tuple of downsample block types. mid_block_type (`str`, *optional*, defaults to `"UNetMidBlock2D"`): Block type for middle of UNet, it can be either `UNetMidBlock2D` or `UnCLIPUNetMidBlock2D`. up_block_types (`Tuple[str]`, *optional*, defaults to `("AttnUpBlock2D", "AttnUpBlock2D", "AttnUpBlock2D", "UpBlock2D")`): Tuple of upsample block types. block_out_channels (`Tuple[int]`, *optional*, defaults to `(224, 448, 672, 896)`): Tuple of block output channels. layers_per_block (`int`, *optional*, defaults to `2`): The number of layers per block. mid_block_scale_factor (`float`, *optional*, defaults to `1`): The scale factor for the mid block. downsample_padding (`int`, *optional*, defaults to `1`): The padding for the downsample convolution. downsample_type (`str`, *optional*, defaults to `conv`): The downsample type for downsampling layers. Choose between "conv" and "resnet" upsample_type (`str`, *optional*, defaults to `conv`): The upsample type for upsampling layers. Choose between "conv" and "resnet" dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use. act_fn (`str`, *optional*, defaults to `"silu"`): The activation function to use. attention_head_dim (`int`, *optional*, defaults to `8`): The attention head dimension. norm_num_groups (`int`, *optional*, defaults to `32`): The number of groups for normalization. attn_norm_num_groups (`int`, *optional*, defaults to `None`): If set to an integer, a group norm layer will be created in the mid block's [`Attention`] layer with the given number of groups. If left as `None`, the group norm layer will only be created if `resnet_time_scale_shift` is set to `default`, and if created will have `norm_num_groups` groups. norm_eps (`float`, *optional*, defaults to `1e-5`): The epsilon for normalization. resnet_time_scale_shift (`str`, *optional*, defaults to `"default"`): Time scale shift config for ResNet blocks (see [`~models.resnet.ResnetBlock2D`]). Choose from `default` or `scale_shift`. class_embed_type (`str`, *optional*, defaults to `None`): The type of class embedding to use which is ultimately summed with the time embeddings. Choose from `None`, `"timestep"`, or `"identity"`. num_class_embeds (`int`, *optional*, defaults to `None`): Input dimension of the learnable embedding matrix to be projected to `time_embed_dim` when performing class conditioning with `class_embed_type` equal to `None`. NrF positionalrT DownBlock2DAttnDownBlock2Dr&r& AttnUpBlock2Dr(r( UpBlock2Diiir rconvsilu h㈵>default sample_size in_channels out_channelscenter_input_sampletime_embedding_type freq_shiftflip_sin_to_cosdown_block_types.up_block_typesblock_out_channelslayers_per_blockmid_block_scale_factordownsample_paddingdownsample_type upsample_typedropoutact_fnattention_head_dimnorm_num_groupsattn_norm_num_groupsnorm_epsresnet_time_scale_shift add_attentionclass_embed_typenum_class_embedsnum_train_timestepsct||_| ddz}t|t| krt d|d| dt| t|krt d| d|dt j|| ddd |_|d kr(t| dd |_ d| dz}nY|dkr%t| d|||_ | d}n.|dkr(t j || d|_ | d}t|||_ ||t j |||_nD|dkrt|||_n(|dkrt j|||_nd|_t jg|_d|_t jg|_| d}t)|D]a\}}|} | |}|t| dz k}!t+|| | |||! |||||n|| |||}"|j|"bt/| d||||| |||n| d||| |_t1t3| }#|#d}t)| D]\}}$|}%|#|}|#t5|dzt| dz } |t| dz k}!t7|$| dz| ||%||! |||||n||||}&|j|&|}%||nt5| ddzd}'t j| d|'||_t j|_t j| d|dd |_ dS)Nrz\Must provide the same number of `down_block_types` as `up_block_types`. `down_block_types`: z. `up_block_types`: .zbMust provide the same number of `block_out_channels` as `down_block_types`. `block_out_channels`: z. `down_block_types`: r)rr) kernel_sizepaddingfourier)embedding_sizescaler r#learnedtimestepidentityr) num_layersr4r5 temb_channelsadd_downsample resnet_eps resnet_act_fn resnet_groupsrDr?rHr@rB) r4rZrBr\r]output_scale_factorrHrDr^ attn_groupsrI) rYr4r5prev_output_channelrZ add_upsampler\r]r^rDrHrArBr0) num_channels num_groupseps)!super__init__r3len ValueErrornnConv2dconv_inr time_projr Embeddingr time_embeddingclass_embeddingIdentity ModuleList down_blocks mid_block up_blocks enumeraterappendrlistreversedminr GroupNorm conv_norm_outSiLUconv_actconv_out))selfr3r4r5r6r7r8r9r:r;r<r=r>r?r@rArBrCrDrErFrGrHrIrJrKrLtime_embed_dimtimestep_input_dimoutput_channelidown_block_type input_channelis_final_block down_blockreversed_block_out_channels up_block_typerbup_blocknum_groups_out __class__s) r rhzUNet2DModel.__init__\s< &+A.2  C$7$7 7 7goggVdggg  ! " "c*:&;&; ; ;svHss`psss  y.@.CQR\bccc  ) + +6FXYZF[cefffDN!"%7%:!:   L 0 0&'9!'CY?Z?Z>ZY*[*[\\Ii'F 9 6****r)NrrFr#rTr$r'r*r rrr,r,r-r.r/r0Nr1r2TNNN)NT)rrrrr rrintrboolstrfloatrhrrrr __classcell__)rs@r r"r"(s11f>B$)#/ $,t*j.B !()"#%#,-!.2'0"*.*.-17TaTaeCsCx$89:TaTa Ta " Ta ! TaTaTa S/Tac3hTa"#s(OTaTa!&Ta TaTa !Ta"#Ta$%Ta&%SM'Ta()Ta*'sm+Ta,-Ta."%/Ta01Ta2#3-3Ta4#3-5Ta6&c]7TaTaTaTaTaTat04 g+g+ g+ eS01g+u|, g+  g+ |U" # g+g+g+g+g+g+g+g+rr") dataclassesrtypingrrrrtorch.nnrkconfiguration_utilsrr utilsr embeddingsr r rmodeling_utilsrunet_2d_blocksrrrrr"rrr rs@"!!!!!)))))))))) BBBBBBBBPPPPPPPPPP''''''HHHHHHHHHH      :    r+r+r+r+r+*kr+r+r+r+r+r