diff --git a/ip_adapter/attention_processor.py b/ip_adapter/attention_processor.py new file mode 100644 index 0000000..60f6a10 --- /dev/null +++ b/ip_adapter/attention_processor.py @@ -0,0 +1,308 @@ +# modified from https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py +import torch +import torch.nn as nn +import torch.nn.functional as F + +try: + import xformers + import xformers.ops + xformers_available = True +except Exception as e: + xformers_available = False + + + +class RegionControler(object): + def __init__(self) -> None: + self.prompt_image_conditioning = [] +region_control = RegionControler() + + +class AttnProcessor(nn.Module): + r""" + Default processor for performing attention-related computations. + """ + def __init__( + self, + hidden_size=None, + cross_attention_dim=None, + ): + super().__init__() + + def __call__( + self, + attn, + hidden_states, + encoder_hidden_states=None, + attention_mask=None, + temb=None, + ): + residual = hidden_states + + if attn.spatial_norm is not None: + hidden_states = attn.spatial_norm(hidden_states, temb) + + input_ndim = hidden_states.ndim + + if input_ndim == 4: + batch_size, channel, height, width = hidden_states.shape + hidden_states = hidden_states.view(batch_size, channel, height * width).transpose(1, 2) + + batch_size, sequence_length, _ = ( + hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape + ) + attention_mask = attn.prepare_attention_mask(attention_mask, sequence_length, batch_size) + + if attn.group_norm is not None: + hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(1, 2) + + query = attn.to_q(hidden_states) + + if encoder_hidden_states is None: + encoder_hidden_states = hidden_states + elif attn.norm_cross: + encoder_hidden_states = attn.norm_encoder_hidden_states(encoder_hidden_states) + + key = attn.to_k(encoder_hidden_states) + value = attn.to_v(encoder_hidden_states) + + query = attn.head_to_batch_dim(query) + key = attn.head_to_batch_dim(key) + value = attn.head_to_batch_dim(value) + + attention_probs = attn.get_attention_scores(query, key, attention_mask) + hidden_states = torch.bmm(attention_probs, value) + hidden_states = attn.batch_to_head_dim(hidden_states) + + # linear proj + hidden_states = attn.to_out[0](hidden_states) + # dropout + hidden_states = attn.to_out[1](hidden_states) + + if input_ndim == 4: + hidden_states = hidden_states.transpose(-1, -2).reshape(batch_size, channel, height, width) + + if attn.residual_connection: + hidden_states = hidden_states + residual + + hidden_states = hidden_states / attn.rescale_output_factor + + return hidden_states + + +class IPAttnProcessor(nn.Module): + r""" + Attention processor for IP-Adapater. + Args: + hidden_size (`int`): + The hidden size of the attention layer. + cross_attention_dim (`int`): + The number of channels in the `encoder_hidden_states`. + scale (`float`, defaults to 1.0): + the weight scale of image prompt. + num_tokens (`int`, defaults to 4 when do ip_adapter_plus it should be 16): + The context length of the image features. + """ + + def __init__(self, hidden_size, cross_attention_dim=None, scale=1.0, num_tokens=4): + super().__init__() + + self.hidden_size = hidden_size + self.cross_attention_dim = cross_attention_dim + self.scale = scale + self.num_tokens = num_tokens + + self.to_k_ip = nn.Linear(cross_attention_dim or hidden_size, hidden_size, bias=False) + self.to_v_ip = nn.Linear(cross_attention_dim or hidden_size, hidden_size, bias=False) + + def __call__( + self, + attn, + hidden_states, + encoder_hidden_states=None, + attention_mask=None, + temb=None, + ): + residual = hidden_states + + if attn.spatial_norm is not None: + hidden_states = attn.spatial_norm(hidden_states, temb) + + input_ndim = hidden_states.ndim + + if input_ndim == 4: + batch_size, channel, height, width = hidden_states.shape + hidden_states = hidden_states.view(batch_size, channel, height * width).transpose(1, 2) + + batch_size, sequence_length, _ = ( + hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape + ) + attention_mask = attn.prepare_attention_mask(attention_mask, sequence_length, batch_size) + + if attn.group_norm is not None: + hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(1, 2) + + query = attn.to_q(hidden_states) + + if encoder_hidden_states is None: + encoder_hidden_states = hidden_states + else: + # get encoder_hidden_states, ip_hidden_states + end_pos = encoder_hidden_states.shape[1] - self.num_tokens + encoder_hidden_states, ip_hidden_states = encoder_hidden_states[:, :end_pos, :], encoder_hidden_states[:, end_pos:, :] + if attn.norm_cross: + encoder_hidden_states = attn.norm_encoder_hidden_states(encoder_hidden_states) + + key = attn.to_k(encoder_hidden_states) + value = attn.to_v(encoder_hidden_states) + + query = attn.head_to_batch_dim(query) + key = attn.head_to_batch_dim(key) + value = attn.head_to_batch_dim(value) + + if xformers_available: + hidden_states = self._memory_efficient_attention_xformers(query, key, value, attention_mask) + else: + attention_probs = attn.get_attention_scores(query, key, attention_mask) + hidden_states = torch.bmm(attention_probs, value) + hidden_states = attn.batch_to_head_dim(hidden_states) + + # for ip-adapter + ip_key = self.to_k_ip(ip_hidden_states) + ip_value = self.to_v_ip(ip_hidden_states) + + ip_key = attn.head_to_batch_dim(ip_key) + ip_value = attn.head_to_batch_dim(ip_value) + + if xformers_available: + ip_hidden_states = self._memory_efficient_attention_xformers(query, ip_key, ip_value, None) + else: + ip_attention_probs = attn.get_attention_scores(query, ip_key, None) + ip_hidden_states = torch.bmm(ip_attention_probs, ip_value) + ip_hidden_states = attn.batch_to_head_dim(ip_hidden_states) + + # region control + if len(region_control.prompt_image_conditioning) == 1: + region_mask = region_control.prompt_image_conditioning[0].get('region_mask', None) + if region_mask is not None: + h, w = region_mask.shape[:2] + ratio = (h * w / query.shape[1]) ** 0.5 + mask = F.interpolate(region_mask[None, None], scale_factor=1/ratio, mode='nearest').reshape([1, -1, 1]) + else: + mask = torch.ones_like(ip_hidden_states) + ip_hidden_states = ip_hidden_states * mask + + hidden_states = hidden_states + self.scale * ip_hidden_states + + # linear proj + hidden_states = attn.to_out[0](hidden_states) + # dropout + hidden_states = attn.to_out[1](hidden_states) + + if input_ndim == 4: + hidden_states = hidden_states.transpose(-1, -2).reshape(batch_size, channel, height, width) + + if attn.residual_connection: + hidden_states = hidden_states + residual + + hidden_states = hidden_states / attn.rescale_output_factor + + return hidden_states + + + def _memory_efficient_attention_xformers(self, query, key, value, attention_mask): + # TODO attention_mask + query = query.contiguous() + key = key.contiguous() + value = value.contiguous() + hidden_states = xformers.ops.memory_efficient_attention(query, key, value, attn_bias=attention_mask) + # hidden_states = self.reshape_batch_dim_to_heads(hidden_states) + return hidden_states + + +class AttnProcessor2_0(torch.nn.Module): + r""" + Processor for implementing scaled dot-product attention (enabled by default if you're using PyTorch 2.0). + """ + def __init__( + self, + hidden_size=None, + cross_attention_dim=None, + ): + super().__init__() + if not hasattr(F, "scaled_dot_product_attention"): + raise ImportError("AttnProcessor2_0 requires PyTorch 2.0, to use it, please upgrade PyTorch to 2.0.") + + def __call__( + self, + attn, + hidden_states, + encoder_hidden_states=None, + attention_mask=None, + temb=None, + ): + residual = hidden_states + + if attn.spatial_norm is not None: + hidden_states = attn.spatial_norm(hidden_states, temb) + + input_ndim = hidden_states.ndim + + if input_ndim == 4: + batch_size, channel, height, width = hidden_states.shape + hidden_states = hidden_states.view(batch_size, channel, height * width).transpose(1, 2) + + batch_size, sequence_length, _ = ( + hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape + ) + + if attention_mask is not None: + attention_mask = attn.prepare_attention_mask(attention_mask, sequence_length, batch_size) + # scaled_dot_product_attention expects attention_mask shape to be + # (batch, heads, source_length, target_length) + attention_mask = attention_mask.view(batch_size, attn.heads, -1, attention_mask.shape[-1]) + + if attn.group_norm is not None: + hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(1, 2) + + query = attn.to_q(hidden_states) + + if encoder_hidden_states is None: + encoder_hidden_states = hidden_states + elif attn.norm_cross: + encoder_hidden_states = attn.norm_encoder_hidden_states(encoder_hidden_states) + + key = attn.to_k(encoder_hidden_states) + value = attn.to_v(encoder_hidden_states) + + inner_dim = key.shape[-1] + head_dim = inner_dim // attn.heads + + query = query.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2) + + key = key.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2) + value = value.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2) + + # the output of sdp = (batch, num_heads, seq_len, head_dim) + # TODO: add support for attn.scale when we move to Torch 2.1 + hidden_states = F.scaled_dot_product_attention( + query, key, value, attn_mask=attention_mask, dropout_p=0.0, is_causal=False + ) + + hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, attn.heads * head_dim) + hidden_states = hidden_states.to(query.dtype) + + # linear proj + hidden_states = attn.to_out[0](hidden_states) + # dropout + hidden_states = attn.to_out[1](hidden_states) + + if input_ndim == 4: + hidden_states = hidden_states.transpose(-1, -2).reshape(batch_size, channel, height, width) + + if attn.residual_connection: + hidden_states = hidden_states + residual + + hidden_states = hidden_states / attn.rescale_output_factor + + return hidden_states \ No newline at end of file diff --git a/ip_adapter/resampler.py b/ip_adapter/resampler.py new file mode 100644 index 0000000..4521c8c --- /dev/null +++ b/ip_adapter/resampler.py @@ -0,0 +1,121 @@ +# modified from https://github.com/mlfoundations/open_flamingo/blob/main/open_flamingo/src/helpers.py +import math + +import torch +import torch.nn as nn + + +# FFN +def FeedForward(dim, mult=4): + inner_dim = int(dim * mult) + return nn.Sequential( + nn.LayerNorm(dim), + nn.Linear(dim, inner_dim, bias=False), + nn.GELU(), + nn.Linear(inner_dim, dim, bias=False), + ) + + +def reshape_tensor(x, heads): + bs, length, width = x.shape + #(bs, length, width) --> (bs, length, n_heads, dim_per_head) + x = x.view(bs, length, heads, -1) + # (bs, length, n_heads, dim_per_head) --> (bs, n_heads, length, dim_per_head) + x = x.transpose(1, 2) + # (bs, n_heads, length, dim_per_head) --> (bs*n_heads, length, dim_per_head) + x = x.reshape(bs, heads, length, -1) + return x + + +class PerceiverAttention(nn.Module): + def __init__(self, *, dim, dim_head=64, heads=8): + super().__init__() + self.scale = dim_head**-0.5 + self.dim_head = dim_head + self.heads = heads + inner_dim = dim_head * heads + + self.norm1 = nn.LayerNorm(dim) + self.norm2 = nn.LayerNorm(dim) + + self.to_q = nn.Linear(dim, inner_dim, bias=False) + self.to_kv = nn.Linear(dim, inner_dim * 2, bias=False) + self.to_out = nn.Linear(inner_dim, dim, bias=False) + + + def forward(self, x, latents): + """ + Args: + x (torch.Tensor): image features + shape (b, n1, D) + latent (torch.Tensor): latent features + shape (b, n2, D) + """ + x = self.norm1(x) + latents = self.norm2(latents) + + b, l, _ = latents.shape + + q = self.to_q(latents) + kv_input = torch.cat((x, latents), dim=-2) + k, v = self.to_kv(kv_input).chunk(2, dim=-1) + + q = reshape_tensor(q, self.heads) + k = reshape_tensor(k, self.heads) + v = reshape_tensor(v, self.heads) + + # attention + scale = 1 / math.sqrt(math.sqrt(self.dim_head)) + weight = (q * scale) @ (k * scale).transpose(-2, -1) # More stable with f16 than dividing afterwards + weight = torch.softmax(weight.float(), dim=-1).type(weight.dtype) + out = weight @ v + + out = out.permute(0, 2, 1, 3).reshape(b, l, -1) + + return self.to_out(out) + + +class Resampler(nn.Module): + def __init__( + self, + dim=1024, + depth=8, + dim_head=64, + heads=16, + num_queries=8, + embedding_dim=768, + output_dim=1024, + ff_mult=4, + ): + super().__init__() + + self.latents = nn.Parameter(torch.randn(1, num_queries, dim) / dim**0.5) + + self.proj_in = nn.Linear(embedding_dim, dim) + + self.proj_out = nn.Linear(dim, output_dim) + self.norm_out = nn.LayerNorm(output_dim) + + self.layers = nn.ModuleList([]) + for _ in range(depth): + self.layers.append( + nn.ModuleList( + [ + PerceiverAttention(dim=dim, dim_head=dim_head, heads=heads), + FeedForward(dim=dim, mult=ff_mult), + ] + ) + ) + + def forward(self, x): + + latents = self.latents.repeat(x.size(0), 1, 1) + + x = self.proj_in(x) + + for attn, ff in self.layers: + latents = attn(x, latents) + latents + latents = ff(latents) + latents + + latents = self.proj_out(latents) + return self.norm_out(latents) \ No newline at end of file diff --git a/ip_adapter/utils.py b/ip_adapter/utils.py new file mode 100644 index 0000000..9a105f3 --- /dev/null +++ b/ip_adapter/utils.py @@ -0,0 +1,5 @@ +import torch.nn.functional as F + + +def is_torch2_available(): + return hasattr(F, "scaled_dot_product_attention")