# Copyright 2023 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 dataclasses import dataclass from typing import Any, Dict, List, Optional, Tuple, Union import os import json from einops import repeat import torch import torch.nn as nn import torch.nn.functional as F import torch.utils.checkpoint from diffusers.configuration_utils import ConfigMixin, register_to_config from diffusers.loaders import UNet2DConditionLoadersMixin from diffusers.utils import BaseOutput, logging from diffusers.models.attention_processor import AttentionProcessor, AttnProcessor from diffusers.models.embeddings import GaussianFourierProjection, TextTimeEmbedding, TimestepEmbedding, Timesteps from diffusers.models.modeling_utils import ModelMixin from unet_blocks import ( all_modules, get_down_block, get_up_block, get_mid_block, ) from unet_utils import FFInflatedConv3d logger = logging.get_logger(__name__) # pylint: disable=invalid-name @dataclass class UNet3DConditionOutput(BaseOutput): """ Args: sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`): Hidden states conditioned on `encoder_hidden_states` input. Output of last layer of model. """ sample: torch.FloatTensor class AudioUNet3DConditionModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin): r""" UNet2DConditionModel is a conditional 2D UNet model that takes in a noisy sample, conditional state, and a timestep and returns sample shaped output. This model inherits from [`ModelMixin`]. Check the superclass documentation for the generic methods the library implements for all the models (such as downloading or saving, etc.) Parameters: sample_size (`int` or `Tuple[int, int]`, *optional*, defaults to `None`): Height and width of input/output sample. in_channels (`int`, *optional*, defaults to 4): The number of channels in the input sample. out_channels (`int`, *optional*, defaults to 4): The number of channels in the output. center_input_sample (`bool`, *optional*, defaults to `False`): Whether to center the input sample. flip_sin_to_cos (`bool`, *optional*, defaults to `False`): Whether to flip the sin to cos in the time embedding. freq_shift (`int`, *optional*, defaults to 0): The frequency shift to apply to the time embedding. down_block_types (`Tuple[str]`, *optional*, defaults to `("CrossAttnDownBlock2D", "CrossAttnDownBlock2D", "CrossAttnDownBlock2D", "DownBlock2D")`): The tuple of downsample blocks to use. mid_block_type (`str`, *optional*, defaults to `"UNetMidBlock2DCrossAttn"`): The mid block type. Choose from `UNetMidBlock2DCrossAttn` or `UNetMidBlock2DSimpleCrossAttn`, will skip the mid block layer if `None`. up_block_types (`Tuple[str]`, *optional*, defaults to `("UpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D", "CrossAttnUpBlock2D",)`): The tuple of upsample blocks to use. only_cross_attention(`bool` or `Tuple[bool]`, *optional*, default to `False`): Whether to include self-attention in the basic transformer blocks, see [`~models.attention.BasicTransformerBlock`]. block_out_channels (`Tuple[int]`, *optional*, defaults to `(320, 640, 1280, 1280)`): The tuple of output channels for each block. layers_per_block (`int`, *optional*, defaults to 2): The number of layers per block. downsample_padding (`int`, *optional*, defaults to 1): The padding to use for the downsampling convolution. mid_block_scale_factor (`float`, *optional*, defaults to 1.0): The scale factor to use for the mid block. act_fn (`str`, *optional*, defaults to `"silu"`): The activation function to use. norm_num_groups (`int`, *optional*, defaults to 32): The number of groups to use for the normalization. If `None`, it will skip the normalization and activation layers in post-processing norm_eps (`float`, *optional*, defaults to 1e-5): The epsilon to use for the normalization. cross_attention_dim (`int` or `Tuple[int]`, *optional*, defaults to 1280): The dimension of the cross attention features. encoder_hid_dim (`int`, *optional*, defaults to None): If given, `encoder_hidden_states` will be projected from this dimension to `cross_attention_dim`. attention_head_dim (`int`, *optional*, defaults to 8): The dimension of the attention heads. 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"`, `"identity"`, `"projection"`, or `"simple_projection"`. addition_embed_type (`str`, *optional*, defaults to None): Configures an optional embedding which will be summed with the time embeddings. Choose from `None` or "text". "text" will use the `TextTimeEmbedding` layer. 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`. time_embedding_type (`str`, *optional*, default to `positional`): The type of position embedding to use for timesteps. Choose from `positional` or `fourier`. time_embedding_dim (`int`, *optional*, default to `None`): An optional override for the dimension of the projected time embedding. time_embedding_act_fn (`str`, *optional*, default to `None`): Optional activation function to use on the time embeddings only one time before they as passed to the rest of the unet. Choose from `silu`, `mish`, `gelu`, and `swish`. timestep_post_act (`str, *optional*, default to `None`): The second activation function to use in timestep embedding. Choose from `silu`, `mish` and `gelu`. time_cond_proj_dim (`int`, *optional*, default to `None`): The dimension of `cond_proj` layer in timestep embedding. conv_in_kernel (`int`, *optional*, default to `3`): The kernel size of `conv_in` layer. conv_out_kernel (`int`, *optional*, default to `3`): The kernel size of `conv_out` layer. projection_class_embeddings_input_dim (`int`, *optional*): The dimension of the `class_labels` input when using the "projection" `class_embed_type`. Required when using the "projection" `class_embed_type`. class_embeddings_concat (`bool`, *optional*, defaults to `False`): Whether to concatenate the time embeddings with the class embeddings. mid_block_only_cross_attention (`bool`, *optional*, defaults to `None`): Whether to use cross attention with the mid block when using the `UNetMidBlock2DSimpleCrossAttn`. If `only_cross_attention` is given as a single boolean and `mid_block_only_cross_attention` is None, the `only_cross_attention` value will be used as the value for `mid_block_only_cross_attention`. Else, it will default to `False`. """ _supports_gradient_checkpointing = True @register_to_config def __init__( self, sample_size: Optional[int] = None, in_channels: int = 4, out_channels: int = 4, center_input_sample: bool = False, flip_sin_to_cos: bool = True, freq_shift: int = 0, down_block_types: Tuple[str] = ( "FFSpatioAudioTempCrossAttnDownBlock3D", "FFSpatioAudioTempCrossAttnDownBlock3D", "FFSpatioAudioTempCrossAttnDownBlock3D", "FFSpatioTempResDownBlock3D", ), mid_block_type: Optional[str] = "FFSpatioAudioTempCrossAttnUNetMidBlock3D", up_block_types: Tuple[str] = ( "FFSpatioTempResUpBlock3D", "FFSpatioAudioTempCrossAttnUpBlock3D", "FFSpatioAudioTempCrossAttnUpBlock3D", "FFSpatioAudioTempCrossAttnUpBlock3D" ), only_cross_attention: Union[bool, Tuple[bool]] = False, block_out_channels: Tuple[int] = (320, 640, 1280, 1280), layers_per_block: Union[int, Tuple[int]] = 2, downsample_padding: int = 1, mid_block_scale_factor: float = 1, act_fn: str = "silu", norm_num_groups: Optional[int] = 32, norm_eps: float = 1e-5, cross_attention_dim: Union[int, Tuple[int]] = 1280, encoder_hid_dim: Optional[int] = None, attention_head_dim: Union[int, Tuple[int]] = 8, dual_cross_attention: bool = False, use_linear_projection: bool = False, class_embed_type: Optional[str] = None, addition_embed_type: Optional[str] = None, num_class_embeds: Optional[int] = None, upcast_attention: bool = False, resnet_time_scale_shift: str = "default", resnet_skip_time_act: bool = False, resnet_out_scale_factor: int = 1.0, time_embedding_type: str = "positional", time_embedding_dim: Optional[int] = None, time_embedding_act_fn: Optional[str] = None, timestep_post_act: Optional[str] = None, time_cond_proj_dim: Optional[int] = None, conv_in_kernel: int = 3, conv_out_kernel: int = 3, projection_class_embeddings_input_dim: Optional[int] = None, class_embeddings_concat: bool = False, mid_block_only_cross_attention: Optional[bool] = None, cross_attention_norm: Optional[str] = None, addition_embed_type_num_heads=64, audio_cross_attention_dim: int = 768, ): super().__init__() self.sample_size = sample_size # Check inputs if len(down_block_types) != len(up_block_types): raise ValueError( f"Must provide the same number of `down_block_types` as `up_block_types`. `down_block_types`: {down_block_types}. `up_block_types`: {up_block_types}." ) if len(block_out_channels) != len(down_block_types): raise ValueError( f"Must provide the same number of `block_out_channels` as `down_block_types`. `block_out_channels`: {block_out_channels}. `down_block_types`: {down_block_types}." ) if not isinstance(only_cross_attention, bool) and len(only_cross_attention) != len(down_block_types): raise ValueError( f"Must provide the same number of `only_cross_attention` as `down_block_types`. `only_cross_attention`: {only_cross_attention}. `down_block_types`: {down_block_types}." ) if not isinstance(attention_head_dim, int) and len(attention_head_dim) != len(down_block_types): raise ValueError( f"Must provide the same number of `attention_head_dim` as `down_block_types`. `attention_head_dim`: {attention_head_dim}. `down_block_types`: {down_block_types}." ) if isinstance(cross_attention_dim, list) and len(cross_attention_dim) != len(down_block_types): raise ValueError( f"Must provide the same number of `cross_attention_dim` as `down_block_types`. `cross_attention_dim`: {cross_attention_dim}. `down_block_types`: {down_block_types}." ) if not isinstance(layers_per_block, int) and len(layers_per_block) != len(down_block_types): raise ValueError( f"Must provide the same number of `layers_per_block` as `down_block_types`. `layers_per_block`: {layers_per_block}. `down_block_types`: {down_block_types}." ) # input conv_in_padding = (conv_in_kernel - 1) // 2 self.conv_in = FFInflatedConv3d( in_channels, block_out_channels[0], kernel_size=conv_in_kernel, padding=conv_in_padding ) # time if time_embedding_type == "fourier": time_embed_dim = time_embedding_dim or block_out_channels[0] * 2 if time_embed_dim % 2 != 0: raise ValueError(f"`time_embed_dim` should be divisible by 2, but is {time_embed_dim}.") self.time_proj = GaussianFourierProjection( time_embed_dim // 2, set_W_to_weight=False, log=False, flip_sin_to_cos=flip_sin_to_cos ) timestep_input_dim = time_embed_dim elif time_embedding_type == "positional": time_embed_dim = time_embedding_dim or block_out_channels[0] * 4 self.time_proj = Timesteps(block_out_channels[0], flip_sin_to_cos, freq_shift) timestep_input_dim = block_out_channels[0] else: raise ValueError( f"{time_embedding_type} does not exist. Please make sure to use one of `fourier` or `positional`." ) self.time_embedding = TimestepEmbedding( timestep_input_dim, time_embed_dim, act_fn=act_fn, post_act_fn=timestep_post_act, cond_proj_dim=time_cond_proj_dim, ) if encoder_hid_dim is not None: self.encoder_hid_proj = nn.Linear(encoder_hid_dim, cross_attention_dim) else: self.encoder_hid_proj = None # class embedding if class_embed_type is None and num_class_embeds is not None: self.class_embedding = nn.Embedding(num_class_embeds, time_embed_dim) elif class_embed_type == "timestep": self.class_embedding = TimestepEmbedding(timestep_input_dim, time_embed_dim, act_fn=act_fn) elif class_embed_type == "identity": self.class_embedding = nn.Identity(time_embed_dim, time_embed_dim) elif class_embed_type == "projection": if projection_class_embeddings_input_dim is None: raise ValueError( "`class_embed_type`: 'projection' requires `projection_class_embeddings_input_dim` be set" ) # The projection `class_embed_type` is the same as the timestep `class_embed_type` except # 1. the `class_labels` inputs are not first converted to sinusoidal embeddings # 2. it projects from an arbitrary input dimension. # # Note that `TimestepEmbedding` is quite general, being mainly linear layers and activations. # When used for embedding actual timesteps, the timesteps are first converted to sinusoidal embeddings. # As a result, `TimestepEmbedding` can be passed arbitrary vectors. self.class_embedding = TimestepEmbedding(projection_class_embeddings_input_dim, time_embed_dim) elif class_embed_type == "simple_projection": if projection_class_embeddings_input_dim is None: raise ValueError( "`class_embed_type`: 'simple_projection' requires `projection_class_embeddings_input_dim` be set" ) self.class_embedding = nn.Linear(projection_class_embeddings_input_dim, time_embed_dim) else: self.class_embedding = None if addition_embed_type == "text": if encoder_hid_dim is not None: text_time_embedding_from_dim = encoder_hid_dim else: text_time_embedding_from_dim = cross_attention_dim self.add_embedding = TextTimeEmbedding( text_time_embedding_from_dim, time_embed_dim, num_heads=addition_embed_type_num_heads ) elif addition_embed_type is not None: raise ValueError(f"addition_embed_type: {addition_embed_type} must be None or 'text'.") if time_embedding_act_fn is None: self.time_embed_act = None elif time_embedding_act_fn == "swish": self.time_embed_act = lambda x: F.silu(x) elif time_embedding_act_fn == "mish": self.time_embed_act = nn.Mish() elif time_embedding_act_fn == "silu": self.time_embed_act = nn.SiLU() elif time_embedding_act_fn == "gelu": self.time_embed_act = nn.GELU() else: raise ValueError(f"Unsupported activation function: {time_embedding_act_fn}") self.down_blocks = nn.ModuleList([]) self.up_blocks = nn.ModuleList([]) if isinstance(only_cross_attention, bool): if mid_block_only_cross_attention is None: mid_block_only_cross_attention = only_cross_attention only_cross_attention = [only_cross_attention] * len(down_block_types) if mid_block_only_cross_attention is None: mid_block_only_cross_attention = False if isinstance(attention_head_dim, int): attention_head_dim = (attention_head_dim,) * len(down_block_types) if isinstance(cross_attention_dim, int): cross_attention_dim = (cross_attention_dim,) * len(down_block_types) if isinstance(layers_per_block, int): layers_per_block = [layers_per_block] * len(down_block_types) if class_embeddings_concat: # The time embeddings are concatenated with the class embeddings. The dimension of the # time embeddings passed to the down, middle, and up blocks is twice the dimension of the # regular time embeddings blocks_time_embed_dim = time_embed_dim * 2 else: blocks_time_embed_dim = time_embed_dim # down output_channel = block_out_channels[0] for i, down_block_type in enumerate(down_block_types): input_channel = output_channel output_channel = block_out_channels[i] is_final_block = i == len(block_out_channels) - 1 down_block = get_down_block( down_block_type, num_layers=layers_per_block[i], in_channels=input_channel, out_channels=output_channel, temb_channels=blocks_time_embed_dim, add_downsample=not is_final_block, resnet_eps=norm_eps, resnet_act_fn=act_fn, resnet_groups=norm_num_groups, cross_attention_dim=cross_attention_dim[i], attn_num_head_channels=attention_head_dim[i], downsample_padding=downsample_padding, dual_cross_attention=dual_cross_attention, use_linear_projection=use_linear_projection, only_cross_attention=only_cross_attention[i], upcast_attention=upcast_attention, resnet_time_scale_shift=resnet_time_scale_shift, audio_cross_attention_dim=audio_cross_attention_dim ) self.down_blocks.append(down_block) # mid if mid_block_type is None: self.mid_block = None else: self.mid_block = get_mid_block( mid_block_type=mid_block_type, in_channels=block_out_channels[-1], temb_channels=blocks_time_embed_dim, resnet_eps=norm_eps, resnet_act_fn=act_fn, output_scale_factor=mid_block_scale_factor, resnet_time_scale_shift=resnet_time_scale_shift, cross_attention_dim=cross_attention_dim[-1], attn_num_head_channels=attention_head_dim[-1], resnet_groups=norm_num_groups, dual_cross_attention=dual_cross_attention, use_linear_projection=use_linear_projection, upcast_attention=upcast_attention, audio_cross_attention_dim=audio_cross_attention_dim ) # count how many layers upsample the images self.num_upsamplers = 0 # up reversed_block_out_channels = list(reversed(block_out_channels)) reversed_attention_head_dim = list(reversed(attention_head_dim)) reversed_layers_per_block = list(reversed(layers_per_block)) reversed_cross_attention_dim = list(reversed(cross_attention_dim)) only_cross_attention = list(reversed(only_cross_attention)) output_channel = reversed_block_out_channels[0] for i, up_block_type in enumerate(up_block_types): is_final_block = i == len(block_out_channels) - 1 prev_output_channel = output_channel output_channel = reversed_block_out_channels[i] input_channel = reversed_block_out_channels[min(i + 1, len(block_out_channels) - 1)] # add upsample block for all BUT final layer if not is_final_block: add_upsample = True self.num_upsamplers += 1 else: add_upsample = False up_block = get_up_block( up_block_type, num_layers=reversed_layers_per_block[i] + 1, in_channels=input_channel, out_channels=output_channel, prev_output_channel=prev_output_channel, temb_channels=blocks_time_embed_dim, add_upsample=add_upsample, resnet_eps=norm_eps, resnet_act_fn=act_fn, resnet_groups=norm_num_groups, cross_attention_dim=reversed_cross_attention_dim[i], attn_num_head_channels=reversed_attention_head_dim[i], dual_cross_attention=dual_cross_attention, use_linear_projection=use_linear_projection, only_cross_attention=only_cross_attention[i], upcast_attention=upcast_attention, resnet_time_scale_shift=resnet_time_scale_shift, audio_cross_attention_dim=audio_cross_attention_dim ) self.up_blocks.append(up_block) # out if norm_num_groups is not None: self.conv_norm_out = nn.GroupNorm( num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=norm_eps ) if act_fn == "swish": self.conv_act = lambda x: F.silu(x) elif act_fn == "mish": self.conv_act = nn.Mish() elif act_fn == "silu": self.conv_act = nn.SiLU() elif act_fn == "gelu": self.conv_act = nn.GELU() else: raise ValueError(f"Unsupported activation function: {act_fn}") else: self.conv_norm_out = None self.conv_act = None conv_out_padding = (conv_out_kernel - 1) // 2 self.conv_out = FFInflatedConv3d( block_out_channels[0], out_channels, kernel_size=conv_out_kernel, padding=conv_out_padding ) @property def attn_processors(self) -> Dict[str, AttentionProcessor]: r""" Returns: `dict` of attention processors: A dictionary containing all attention processors used in the model with indexed by its weight name. """ # set recursively processors = {} def fn_recursive_add_processors(name: str, module: torch.nn.Module, processors: Dict[str, AttentionProcessor]): if hasattr(module, "set_processor"): processors[f"{name}.processor"] = module.processor for sub_name, child in module.named_children(): fn_recursive_add_processors(f"{name}.{sub_name}", child, processors) return processors for name, module in self.named_children(): fn_recursive_add_processors(name, module, processors) return processors def set_attn_processor(self, processor: Union[AttentionProcessor, Dict[str, AttentionProcessor]]): r""" Parameters: `processor (`dict` of `AttentionProcessor` or `AttentionProcessor`): The instantiated processor class or a dictionary of processor classes that will be set as the processor of **all** `Attention` layers. In case `processor` is a dict, the key needs to define the path to the corresponding cross attention processor. This is strongly recommended when setting trainable attention processors.: """ count = len(self.attn_processors.keys()) if isinstance(processor, dict) and len(processor) != count: raise ValueError( f"A dict of processors was passed, but the number of processors {len(processor)} does not match the" f" number of attention layers: {count}. Please make sure to pass {count} processor classes." ) def fn_recursive_attn_processor(name: str, module: torch.nn.Module, processor): if hasattr(module, "set_processor"): if not isinstance(processor, dict): module.set_processor(processor) else: module.set_processor(processor.pop(f"{name}.processor")) for sub_name, child in module.named_children(): fn_recursive_attn_processor(f"{name}.{sub_name}", child, processor) for name, module in self.named_children(): fn_recursive_attn_processor(name, module, processor) def set_default_attn_processor(self): """ Disables custom attention processors and sets the default attention implementation. """ self.set_attn_processor(AttnProcessor()) def set_attention_slice(self, slice_size): r""" Enable sliced attention computation. When this option is enabled, the attention module will split the input tensor in slices, to compute attention in several steps. This is useful to save some memory in exchange for a small speed decrease. Args: slice_size (`str` or `int` or `list(int)`, *optional*, defaults to `"auto"`): When `"auto"`, halves the input to the attention heads, so attention will be computed in two steps. If `"max"`, maximum amount of memory will be saved by running only one slice at a time. If a number is provided, uses as many slices as `attention_head_dim // slice_size`. In this case, `attention_head_dim` must be a multiple of `slice_size`. """ sliceable_head_dims = [] def fn_recursive_retrieve_sliceable_dims(module: torch.nn.Module): if hasattr(module, "set_attention_slice"): sliceable_head_dims.append(module.sliceable_head_dim) for child in module.children(): fn_recursive_retrieve_sliceable_dims(child) # retrieve number of attention layers for module in self.children(): fn_recursive_retrieve_sliceable_dims(module) num_sliceable_layers = len(sliceable_head_dims) if slice_size == "auto": # half the attention head size is usually a good trade-off between # speed and memory slice_size = [dim // 2 for dim in sliceable_head_dims] elif slice_size == "max": # make smallest slice possible slice_size = num_sliceable_layers * [1] slice_size = num_sliceable_layers * [slice_size] if not isinstance(slice_size, list) else slice_size if len(slice_size) != len(sliceable_head_dims): raise ValueError( f"You have provided {len(slice_size)}, but {self.config} has {len(sliceable_head_dims)} different" f" attention layers. Make sure to match `len(slice_size)` to be {len(sliceable_head_dims)}." ) for i in range(len(slice_size)): size = slice_size[i] dim = sliceable_head_dims[i] if size is not None and size > dim: raise ValueError(f"size {size} has to be smaller or equal to {dim}.") # Recursively walk through all the children. # Any children which exposes the set_attention_slice method # gets the message def fn_recursive_set_attention_slice(module: torch.nn.Module, slice_size: List[int]): if hasattr(module, "set_attention_slice"): module.set_attention_slice(slice_size.pop()) for child in module.children(): fn_recursive_set_attention_slice(child, slice_size) reversed_slice_size = list(reversed(slice_size)) for module in self.children(): fn_recursive_set_attention_slice(module, reversed_slice_size) def _set_gradient_checkpointing(self, module, value=False): if isinstance(module, tuple(all_modules)): module.gradient_checkpointing = value def forward( self, sample: torch.FloatTensor, timestep: Union[torch.Tensor, float, int], encoder_hidden_states: torch.Tensor, audio_encoder_hidden_states: Optional[torch.Tensor] = None, class_labels: Optional[torch.Tensor] = None, timestep_cond: Optional[torch.Tensor] = None, attention_mask: Optional[torch.Tensor] = None, audio_attention_mask: Optional[torch.Tensor] = None, cross_attention_kwargs: Optional[Dict[str, Any]] = None, down_block_additional_residuals: Optional[Tuple[torch.Tensor]] = None, mid_block_additional_residual: Optional[torch.Tensor] = None, return_dict: bool = True, ) -> Union[UNet3DConditionOutput, Tuple]: r""" Args: sample (`torch.FloatTensor`): (batch, channel, frame, height, width) noisy inputs tensor timestep (`torch.FloatTensor` or `float` or `int`): (batch) timesteps encoder_hidden_states (`torch.FloatTensor`): (batch, sequence_length, feature_dim) encoder hidden states return_dict (`bool`, *optional*, defaults to `True`): Whether or not to return a [`models.unet_2d_condition.UNet2DConditionOutput`] instead of a plain tuple. cross_attention_kwargs (`dict`, *optional*): A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under `self.processor` in [diffusers.cross_attention](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/cross_attention.py). Returns: [`~models.unet_2d_condition.UNet2DConditionOutput`] or `tuple`: [`~models.unet_2d_condition.UNet2DConditionOutput`] if `return_dict` is True, otherwise a `tuple`. When returning a tuple, the first element is the sample tensor. """ assert sample.ndim == 5, sample.size() video_length = sample.shape[2] # By default samples have to be AT least a multiple of the overall upsampling factor. # The overall upsampling factor is equal to 2 ** (# num of upsampling layers). # However, the upsampling interpolation output size can be forced to fit any upsampling size # on the fly if necessary. default_overall_up_factor = 2 ** self.num_upsamplers # upsample size should be forwarded when sample is not a multiple of `default_overall_up_factor` forward_upsample_size = False upsample_size = None if any(s % default_overall_up_factor != 0 for s in sample.shape[-2:]): logger.info("Forward upsample size to force interpolation output size.") forward_upsample_size = True # prepare attention_mask if attention_mask is not None: attention_mask = (1 - attention_mask.to(sample.dtype)) * -10000.0 attention_mask = attention_mask.unsqueeze(1) # 0. center input if necessary if self.config.center_input_sample: sample = 2 * sample - 1.0 # 1. time timesteps = timestep if not torch.is_tensor(timesteps): # TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can # This would be a good case for the `match` statement (Python 3.10+) is_mps = sample.device.type == "mps" if isinstance(timestep, float): dtype = torch.float32 if is_mps else torch.float64 else: dtype = torch.int32 if is_mps else torch.int64 timesteps = torch.tensor([timesteps], dtype=dtype, device=sample.device) elif len(timesteps.shape) == 0: timesteps = timesteps[None].to(sample.device) # broadcast to batch dimension in a way that's compatible with ONNX/Core ML timesteps = timesteps.expand(sample.shape[0]) t_emb = self.time_proj(timesteps) # `Timesteps` does not contain any weights and will always return f32 tensors # but time_embedding might actually be running in fp16. so we need to cast here. # there might be better ways to encapsulate this. t_emb = t_emb.to(dtype=self.dtype) emb = self.time_embedding(t_emb, timestep_cond) emb = repeat(emb, "b c -> b f c", f=video_length) if self.class_embedding is not None: if class_labels is None: raise ValueError("class_labels should be provided when num_class_embeds > 0") if self.config.class_embed_type == "timestep": class_labels = self.time_proj(class_labels) # `Timesteps` does not contain any weights and will always return f32 tensors # there might be better ways to encapsulate this. class_labels = class_labels.to(dtype=sample.dtype) class_emb = self.class_embedding(class_labels).to(dtype=self.dtype) if self.config.class_embeddings_concat: emb = torch.cat([emb, class_emb], dim=-1) else: emb = emb + class_emb if self.config.addition_embed_type == "text": aug_emb = self.add_embedding(encoder_hidden_states) emb = emb + aug_emb if self.time_embed_act is not None: emb = self.time_embed_act(emb) if self.encoder_hid_proj is not None: encoder_hidden_states = self.encoder_hid_proj(encoder_hidden_states) # 2. pre-process sample = self.conv_in(sample) # 3. down down_block_res_samples = (sample,) for downsample_block in self.down_blocks: if hasattr(downsample_block, "has_cross_attention") and downsample_block.has_cross_attention: sample, res_samples = downsample_block( hidden_states=sample, temb=emb, encoder_hidden_states=encoder_hidden_states, audio_encoder_hidden_states=audio_encoder_hidden_states, attention_mask=attention_mask, audio_attention_mask=audio_attention_mask, cross_attention_kwargs=cross_attention_kwargs, ) else: sample, res_samples = downsample_block( hidden_states=sample, temb=emb ) down_block_res_samples += res_samples if down_block_additional_residuals is not None: new_down_block_res_samples = () for down_block_res_sample, down_block_additional_residual in zip( down_block_res_samples, down_block_additional_residuals ): down_block_res_sample = down_block_res_sample + down_block_additional_residual new_down_block_res_samples += (down_block_res_sample,) down_block_res_samples = new_down_block_res_samples # 4. mid if self.mid_block is not None: sample = self.mid_block( sample, emb, encoder_hidden_states=encoder_hidden_states, audio_encoder_hidden_states=audio_encoder_hidden_states, attention_mask=attention_mask, audio_attention_mask=audio_attention_mask, cross_attention_kwargs=cross_attention_kwargs, ) if mid_block_additional_residual is not None: sample = sample + mid_block_additional_residual # 5. up for i, upsample_block in enumerate(self.up_blocks): is_final_block = i == len(self.up_blocks) - 1 res_samples = down_block_res_samples[-len(upsample_block.resnets):] down_block_res_samples = down_block_res_samples[: -len(upsample_block.resnets)] # if we have not reached the final block and need to forward the # upsample size, we do it here if not is_final_block and forward_upsample_size: upsample_size = down_block_res_samples[-1].shape[2:] if hasattr(upsample_block, "has_cross_attention") and upsample_block.has_cross_attention: sample = upsample_block( hidden_states=sample, temb=emb, res_hidden_states_tuple=res_samples, encoder_hidden_states=encoder_hidden_states, audio_encoder_hidden_states=audio_encoder_hidden_states, cross_attention_kwargs=cross_attention_kwargs, upsample_size=upsample_size, attention_mask=attention_mask, audio_attention_mask=audio_attention_mask, ) else: sample = upsample_block( hidden_states=sample, temb=emb, res_hidden_states_tuple=res_samples, upsample_size=upsample_size ) # 6. post-process if self.conv_norm_out: sample = self.conv_norm_out(sample) sample = self.conv_act(sample) sample = self.conv_out(sample) if not return_dict: return (sample,) return UNet3DConditionOutput(sample=sample) @classmethod def from_pretrained_2d(cls, config3d, pretrained_model_path, subfolder=None): # 1. Build 3D config from pretrained 2D config if subfolder is not None: pretrained_model_path = os.path.join(pretrained_model_path, subfolder) config2d_file = os.path.join(pretrained_model_path, 'config.json') assert os.path.isfile(config2d_file), f"{config2d_file} does not exist" with open(config2d_file, "r") as f: config2d = json.load(f) config2d["_class_name"] = cls.__name__ config2d["down_block_types"] = tuple(config3d["down_block_types"]) config2d["up_block_types"] = tuple(config3d["up_block_types"]) config2d["mid_block_type"] = config3d["mid_block_type"] if "cross_attention_dim" in config3d: config2d["cross_attention_dim"] = config3d["cross_attention_dim"] if "audio_cross_attention_dim" in config3d: config2d["audio_cross_attention_dim"] = config3d[ "audio_cross_attention_dim"] # 2. Build 3D model from updated 3D config model = cls.from_config(config2d) # 3. Load in weights from pretrained 2D nets from diffusers.utils import WEIGHTS_NAME model2d_file = os.path.join(pretrained_model_path, WEIGHTS_NAME) assert os.path.isfile(model2d_file), f"{model2d_file} does not exist" pretrained_2d_state_dict = torch.load(model2d_file, map_location="cpu") # Add new 3D weights into pretrained 2d state_dict, to be compatible with 3D model for k, v in model.state_dict().items(): # all '_temp' temporal weights are initialized by pretrained 2D models if '_temp' in k: pretrained_2d_state_dict.update({k: v}) # add new weights into pretrained 2D state_dict elif k not in pretrained_2d_state_dict: pretrained_2d_state_dict.update({k: v}) # if weights has different shape, replace it elif pretrained_2d_state_dict[k].shape != v.shape: pretrained_2d_state_dict.update({k: v}) model.load_state_dict(pretrained_2d_state_dict) return model