Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/magic-research/magic-animate
/ types & classes
Types & classes
57 in github.com/magic-research/magic-animate
⨍
Functions
239
◇
Types & classes
57
↓ 12 callers
Class
TimestepEmbedding
magicanimate/models/embeddings.py:161
↓ 9 callers
Class
InflatedConv3d
magicanimate/models/resnet.py:30
↓ 8 callers
Class
ReferenceAttentionControl
magicanimate/models/mutual_self_attention.py:128
↓ 6 callers
Class
Identity
r"""A placeholder identity operator that is argument-insensitive. Args: args: any argument (unused) kwargs: any keyword argument
magicanimate/models/appearance_encoder.py:63
↓ 6 callers
Class
ResnetBlock3D
magicanimate/models/resnet.py:113
↓ 6 callers
Class
Timesteps
magicanimate/models/embeddings.py:221
↓ 5 callers
Class
AdaLayerNorm
Norm layer modified to incorporate timestep embeddings.
magicanimate/models/orig_attention.py:846
↓ 5 callers
Class
CrossAttention
r""" A cross attention layer. Parameters: query_dim (`int`): The number of channels in the query. cross_attention_dim (`int`,
magicanimate/models/orig_attention.py:516
↓ 4 callers
Class
VideoReader
Simple wrapper around PyAV that exposes a few useful functions for dealing with video reading. PyAV is a pythonic binding for the ffmpeg libr
magicanimate/utils/videoreader.py:31
↓ 3 callers
Class
AnimationPipeline
magicanimate/pipelines/pipeline_animation.py:73
↓ 3 callers
Class
FeedForward
r""" A feed-forward layer. Parameters: dim (`int`): The number of channels in the input. dim_out (`int`, *optional*): The num
magicanimate/models/orig_attention.py:739
↓ 3 callers
Class
Transformer3DModel
magicanimate/models/attention.py:48
↓ 3 callers
Class
_LoRACompatibleLinear
A Linear layer that can be used with LoRA.
magicanimate/models/appearance_encoder.py:91
↓ 2 callers
Class
Downsample3D
magicanimate/models/resnet.py:87
↓ 2 callers
Class
MagicAnimate
demo/animate_dist.py:42
↓ 2 callers
Class
Transformer2DModelOutput
Args: sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` or `(batch size, num_vector_embeds - 1, num_latent
magicanimate/models/orig_attention.py:36
↓ 2 callers
Class
UNetMidBlock3DCrossAttn
magicanimate/models/unet_3d_blocks.py:181
↓ 2 callers
Class
Upsample3D
magicanimate/models/resnet.py:41
↓ 1 callers
Class
AnimationPipelineOutput
magicanimate/pipelines/pipeline_animation.py:69
↓ 1 callers
Class
ApproximateGELU
The approximate form of Gaussian Error Linear Unit (GELU) For more details, see section 2: https://arxiv.org/abs/1606.08415
magicanimate/models/orig_attention.py:830
↓ 1 callers
Class
BasicTransformerBlock
r""" A basic Transformer block. Parameters: dim (`int`): The number of channels in the input and output. num_attention_heads
magicanimate/models/orig_attention.py:388
↓ 1 callers
Class
BasicTransformerBlock
magicanimate/models/attention.py:164
↓ 1 callers
Class
ControlNetConditioningEmbedding
Quoting from https://arxiv.org/abs/2302.05543: "Stable Diffusion uses a pre-processing method similar to VQ-GAN [11] to convert the entire da
magicanimate/models/controlnet.py:49
↓ 1 callers
Class
ControlNetOutput
magicanimate/models/controlnet.py:44
↓ 1 callers
Class
CrossAttnDownBlock3D
magicanimate/models/unet_3d_blocks.py:286
↓ 1 callers
Class
CrossAttnUpBlock3D
magicanimate/models/unet_3d_blocks.py:522
↓ 1 callers
Class
DownBlock3D
magicanimate/models/unet_3d_blocks.py:426
↓ 1 callers
Class
GEGLU
r""" A variant of the gated linear unit activation function from https://arxiv.org/abs/2002.05202. Parameters: dim_in (`int`): The nu
magicanimate/models/orig_attention.py:806
↓ 1 callers
Class
GELU
r""" GELU activation function
magicanimate/models/orig_attention.py:784
↓ 1 callers
Class
GaussianFourierProjection
Gaussian Fourier embeddings for noise levels.
magicanimate/models/embeddings.py:238
↓ 1 callers
Class
ImagePositionalEmbeddings
Converts latent image classes into vector embeddings. Sums the vector embeddings with positional embeddings for the height and width of the l
magicanimate/models/embeddings.py:268
↓ 1 callers
Class
LabelEmbedding
Embeds class labels into vector representations. Also handles label dropout for classifier-free guidance. Args: num_classes (`int`):
magicanimate/models/embeddings.py:333
↓ 1 callers
Class
MagicAnimate
demo/animate.py:45
↓ 1 callers
Class
Mish
magicanimate/models/resnet.py:210
↓ 1 callers
Class
PositionalEncoding
magicanimate/models/motion_module.py:230
↓ 1 callers
Class
TemporalTransformer3DModel
magicanimate/models/motion_module.py:90
↓ 1 callers
Class
TemporalTransformerBlock
magicanimate/models/motion_module.py:166
↓ 1 callers
Class
Transformer2DModel
Transformer model for image-like data. Takes either discrete (classes of vector embeddings) or continuous (actual embeddings) inputs. Wh
magicanimate/models/orig_attention.py:54
↓ 1 callers
Class
Transformer3DModelOutput
magicanimate/models/attention.py:37
↓ 1 callers
Class
UNet2DConditionOutput
The output of [`UNet2DConditionModel`]. Args: sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)`):
magicanimate/models/appearance_encoder.py:114
↓ 1 callers
Class
UNet3DConditionOutput
magicanimate/models/unet.py:53
↓ 1 callers
Class
UNet3DConditionOutput
magicanimate/models/unet_controlnet.py:50
↓ 1 callers
Class
UpBlock3D
magicanimate/models/unet_3d_blocks.py:665
↓ 1 callers
Class
VanillaTemporalModule
magicanimate/models/motion_module.py:53
↓ 1 callers
Class
VersatileAttention
magicanimate/models/motion_module.py:251
Class
AppearanceEncoderModel
r""" A conditional 2D UNet model that takes a noisy sample, conditional state, and a timestep and returns a sample shaped output. This mo
magicanimate/models/appearance_encoder.py:126
Class
AttentionBase
magicanimate/models/mutual_self_attention.py:24
Class
AttentionBlock
An attention block that allows spatial positions to attend to each other. Originally ported from here, but adapted to the N-d case. https
magicanimate/models/orig_attention.py:253
Class
CombinedTimestepLabelEmbeddings
magicanimate/models/embeddings.py:369
Class
ControlNetModel
magicanimate/models/controlnet.py:94
Class
DualTransformer2DModel
Dual transformer wrapper that combines two `Transformer2DModel`s for mixed inference. Parameters: num_attention_heads (`int`, *optio
magicanimate/models/orig_attention.py:865
Class
MutualSelfAttentionControl
magicanimate/models/mutual_self_attention.py:53
Class
PatchEmbed
2D Image to Patch Embedding
magicanimate/models/embeddings.py:121
Class
StableDiffusionControlNetReferencePipeline
magicanimate/models/stable_diffusion_controlnet_reference.py:72
Class
TemporalTransformer3DModelOutput
magicanimate/models/motion_module.py:31
Class
UNet3DConditionModel
magicanimate/models/unet.py:57
Class
UNet3DConditionModel
magicanimate/models/unet_controlnet.py:54