from collections import OrderedDict from typing import Tuple, Union import torch import torch.nn.functional as F from torch import nn class Bottleneck(nn.Module): expansion = 4 def __init__(self, inplanes, planes, stride=1): super().__init__() # all conv layers have stride 1. an avgpool is performed after the second convolution when stride > 1 self.conv1 = nn.Conv2d(inplanes, planes, 1, bias=False) self.bn1 = nn.BatchNorm2d(planes) self.conv2 = nn.Conv2d(planes, planes, 3, padding=1, bias=False) self.bn2 = nn.BatchNorm2d(planes) self.avgpool = nn.AvgPool2d(stride) if stride > 1 else nn.Identity() self.conv3 = nn.Conv2d(planes, planes * self.expansion, 1, bias=False) self.bn3 = nn.BatchNorm2d(planes * self.expansion) self.relu = nn.ReLU(inplace=True) self.downsample = None self.stride = stride if stride > 1 or inplanes != planes * Bottleneck.expansion: # downsampling layer is prepended with an avgpool, and the subsequent convolution has stride 1 self.downsample = nn.Sequential(OrderedDict([ ("-1", nn.AvgPool2d(stride)), ("0", nn.Conv2d(inplanes, planes * self.expansion, 1, stride=1, bias=False)), ("1", nn.BatchNorm2d(planes * self.expansion)) ])) def forward(self, x: torch.Tensor): identity = x out = self.relu(self.bn1(self.conv1(x))) out = self.relu(self.bn2(self.conv2(out))) out = self.avgpool(out) out = self.bn3(self.conv3(out)) if self.downsample is not None: identity = self.downsample(x) out += identity out = self.relu(out) return out class AttentionPool2d(nn.Module): def __init__(self, spacial_dim: int, embed_dim: int, num_heads: int, output_dim: int = None): super().__init__() self.positional_embedding = nn.Parameter(torch.randn(spacial_dim ** 2 + 1, embed_dim) / embed_dim ** 0.5) self.k_proj = nn.Linear(embed_dim, embed_dim) self.q_proj = nn.Linear(embed_dim, embed_dim) self.v_proj = nn.Linear(embed_dim, embed_dim) self.c_proj = nn.Linear(embed_dim, output_dim or embed_dim) self.num_heads = num_heads def forward(self, x): x = x.flatten(start_dim=2).permute(2, 0, 1) # NCHW -> (HW)NC x = torch.cat([x.mean(dim=0, keepdim=True), x], dim=0) # (HW+1)NC x = x + self.positional_embedding[:, None, :].to(x.dtype) # (HW+1)NC x, _ = F.multi_head_attention_forward( query=x[:1], key=x, value=x, embed_dim_to_check=x.shape[-1], num_heads=self.num_heads, q_proj_weight=self.q_proj.weight, k_proj_weight=self.k_proj.weight, v_proj_weight=self.v_proj.weight, in_proj_weight=None, in_proj_bias=torch.cat([self.q_proj.bias, self.k_proj.bias, self.v_proj.bias]), bias_k=None, bias_v=None, add_zero_attn=False, dropout_p=0, out_proj_weight=self.c_proj.weight, out_proj_bias=self.c_proj.bias, use_separate_proj_weight=True, training=self.training, need_weights=False ) return x.squeeze(0) class ModifiedResNet(nn.Module): """ A ResNet class that is similar to torchvision's but contains the following changes: - There are now 3 "stem" convolutions as opposed to 1, with an average pool instead of a max pool. - Performs anti-aliasing strided convolutions, where an avgpool is prepended to convolutions with stride > 1 - The final pooling layer is a QKV attention instead of an average pool """ def __init__(self, layers, output_dim, heads, input_resolution=224, width=64): super().__init__() self.output_dim = output_dim self.input_resolution = input_resolution # the 3-layer stem self.conv1 = nn.Conv2d(3, width // 2, kernel_size=3, stride=2, padding=1, bias=False) self.bn1 = nn.BatchNorm2d(width // 2) self.relu1 = nn.ReLU(inplace=True) self.conv2 = nn.Conv2d(width // 2, width // 2, kernel_size=3, padding=1, bias=False) self.bn2 = nn.BatchNorm2d(width // 2) self.relu2 = nn.ReLU(inplace=True) self.conv3 = nn.Conv2d(width // 2, width, kernel_size=3, padding=1, bias=False) self.bn3 = nn.BatchNorm2d(width) self.relu3 = nn.ReLU(inplace=True) self.avgpool = nn.AvgPool2d(2) # residual layers self._inplanes = width # this is a *mutable* variable used during construction self.layer1 = self._make_layer(width, layers[0]) self.layer2 = self._make_layer(width * 2, layers[1], stride=2) self.layer3 = self._make_layer(width * 4, layers[2], stride=2) self.layer4 = self._make_layer(width * 8, layers[3], stride=2) embed_dim = width * 32 # the ResNet feature dimension self.attnpool = AttentionPool2d(input_resolution // 32, embed_dim, heads, output_dim) def _make_layer(self, planes, blocks, stride=1): layers = [Bottleneck(self._inplanes, planes, stride)] self._inplanes = planes * Bottleneck.expansion for _ in range(1, blocks): layers.append(Bottleneck(self._inplanes, planes)) return nn.Sequential(*layers) def forward(self, x): def stem(x): x = self.relu1(self.bn1(self.conv1(x))) x = self.relu2(self.bn2(self.conv2(x))) x = self.relu3(self.bn3(self.conv3(x))) x = self.avgpool(x) return x x = x.type(self.conv1.weight.dtype) x = stem(x) x = self.layer1(x) x = self.layer2(x) x = self.layer3(x) x = self.layer4(x) x = self.attnpool(x) return x class LayerNorm(nn.LayerNorm): """Subclass torch's LayerNorm to handle fp16.""" def forward(self, x: torch.Tensor): orig_type = x.dtype ret = super().forward(x.type(torch.float32)) return ret.type(orig_type) class QuickGELU(nn.Module): def forward(self, x: torch.Tensor): return x * torch.sigmoid(1.702 * x) class ResidualAttentionBlock(nn.Module): def __init__(self, d_model: int, n_head: int, attn_mask: torch.Tensor = None): super().__init__() self.attn = nn.MultiheadAttention(d_model, n_head) self.ln_1 = LayerNorm(d_model) self.mlp = nn.Sequential(OrderedDict([ ("c_fc", nn.Linear(d_model, d_model * 4)), ("gelu", QuickGELU()), ("c_proj", nn.Linear(d_model * 4, d_model)) ])) self.ln_2 = LayerNorm(d_model) self.attn_mask = attn_mask self.mask_pre_mlp = True def attention(self, x: torch.Tensor): self.attn_mask = self.attn_mask.to(dtype=x.dtype, device=x.device) if self.attn_mask is not None else None return self.attn(x, x, x, need_weights=False, attn_mask=self.attn_mask)[0] def forward(self, x: torch.Tensor): x = x + self.attention(self.ln_1(x)) x = x + self.mlp(self.ln_2(x)) return x def forward_dense(self, x: torch.Tensor): y = self.ln_1(x) y = F.linear(y, self.attn.in_proj_weight, self.attn.in_proj_bias) L, N, D = y.shape # L N 3D y = y.reshape(L, N, 3, D // 3).permute(2, 1, 0, 3).reshape(3 * N, L, D // 3) y = F.linear(y, self.attn.out_proj.weight, self.attn.out_proj.bias) q, k, v = y.tensor_split(3, dim=0) #v = v.transpose(1, 0) + x # L N D v = v.transpose(1, 0) + x[:1] # L N D v = v + self.mlp(self.ln_2(v)) return v class Transformer(nn.Module): def __init__(self, width: int, layers: int, heads: int, attn_mask: torch.Tensor = None, prompt_length=0, prompt_depth=0): super().__init__() self.width = width self.layers = layers self.resblocks = nn.Sequential(*[ResidualAttentionBlock(width, heads, attn_mask) for _ in range(layers)]) self.prompt_length = prompt_length self.prompt_depth = prompt_depth self.prompt_tokens = nn.Parameter(torch.zeros(prompt_depth, prompt_length, width)) if prompt_length > 0 else None if self.prompt_tokens is not None: nn.init.xavier_uniform_(self.prompt_tokens) def forward(self, x: torch.Tensor, dense=False): for i, resblock in enumerate(self.resblocks): if self.prompt_length > 0 and i < self.prompt_depth: l = self.prompt_length + 1 if i > 0 else 1 x = torch.cat((x[0:1, :, :], self.prompt_tokens[i].repeat(x.shape[1], 1, 1).permute(1, 0, 2) ,x[l:, :, :])) if i == self.layers - 1 and dense: x = resblock.forward_dense(x) x = torch.cat((x[0:1, :, :], x[self.prompt_length + 1: :, :]), dim=0) else: x = resblock(x) return x class VisualTransformer(nn.Module): def __init__(self, input_resolution: int, patch_size: int, width: int, layers: int, heads: int, output_dim: int, prompt_depth: int, prompt_length: int): super().__init__() self.output_dim = output_dim self.conv1 = nn.Conv2d(in_channels=3, out_channels=width, kernel_size=patch_size, stride=patch_size, bias=False) scale = width ** -0.5 self.class_embedding = nn.Parameter(scale * torch.randn(width)) self.positional_embedding = nn.Parameter(scale * torch.randn((input_resolution // patch_size) ** 2 + 1, width)) self.ln_pre = LayerNorm(width) self.transformer = Transformer(width, layers, heads, prompt_depth=prompt_depth, prompt_length=prompt_length) self.ln_post = LayerNorm(width) self.proj = nn.Parameter(scale * torch.randn(width, output_dim)) self.patch_size = patch_size self.input_resolution = input_resolution def forward(self, x: torch.Tensor, dense=False): x = self.conv1(x) # shape = [*, width, grid, grid] x = x.reshape(x.shape[0], x.shape[1], -1) # shape = [*, width, grid ** 2] x = x.permute(0, 2, 1) # shape = [*, grid ** 2, width] x = torch.cat([self.class_embedding.to(x.dtype) + torch.zeros(x.shape[0], 1, x.shape[-1], dtype=x.dtype, device=x.device), x], dim=1) # shape = [*, grid ** 2 + 1, width] if dense and (x.shape[1] != self.positional_embedding.shape[0]): x = x + self.resized_pos_embed(self.input_resolution, x.shape[1]).to(x.dtype) else: x = x + self.positional_embedding.to(x.dtype) x = self.ln_pre(x) x = x.permute(1, 0, 2) # NLD -> LND x = self.transformer(x, dense) x = x.permute(1, 0, 2) # LND -> NLD if dense: x = self.ln_post(x[:, :, :]) else: x = self.ln_post(x[:, 0, :]) if self.proj is not None: x = x @ self.proj return x def resized_pos_embed(self, in_res, tgt_res, mode="bicubic"): #assert L == (input_resolution // self.patch_size) ** 2 + 1 L, D = self.positional_embedding.shape in_side = in_res // self.patch_size #tgt_side = tgt_res // self.patch_size tgt_side = int((tgt_res - 1) ** 0.5) cls_pos = self.positional_embedding[0].unsqueeze(0) # 1 D pos_embed = self.positional_embedding[1:].reshape(1, in_side, in_side, D).permute(0, 3, 1, 2) # L-1 D -> 1 D S S resized_pos_embed = F.interpolate(pos_embed, size=(tgt_side, tgt_side), mode=mode, align_corners=False,) # 1 D S S -> 1 D S' S' resized_pos_embed = resized_pos_embed.squeeze(0).reshape(D, -1).T # L'-1 D return torch.cat((cls_pos, resized_pos_embed), dim=0) class CLIP(nn.Module): def __init__(self, embed_dim: int, # vision image_resolution: int, vision_layers: Union[Tuple[int, int, int, int], int], vision_width: int, vision_patch_size: int, # text context_length: int, vocab_size: int, transformer_width: int, transformer_heads: int, transformer_layers: int, # prompt prompt_depth: int=0, prompt_length: int=0, ): super().__init__() self.context_length = context_length self.image_resolution = image_resolution if isinstance(vision_layers, (tuple, list)): assert prompt_length == 0 and prompt_depth==0 vision_heads = vision_width * 32 // 64 self.visual = ModifiedResNet( layers=vision_layers, output_dim=embed_dim, heads=vision_heads, input_resolution=image_resolution, width=vision_width ) else: vision_heads = vision_width // 64 self.visual = VisualTransformer( input_resolution=image_resolution, patch_size=vision_patch_size, width=vision_width, layers=vision_layers, heads=vision_heads, output_dim=embed_dim, prompt_depth=prompt_depth, prompt_length=prompt_length, ) self.transformer = Transformer( width=transformer_width, layers=transformer_layers, heads=transformer_heads, attn_mask=self.build_attention_mask() ) self.vocab_size = vocab_size self.token_embedding = nn.Embedding(vocab_size, transformer_width) self.positional_embedding = nn.Parameter(torch.empty(self.context_length, transformer_width)) self.ln_final = LayerNorm(transformer_width) self.text_projection = nn.Parameter(torch.empty(transformer_width, embed_dim)) self.logit_scale = nn.Parameter(torch.ones([])) def build_attention_mask(self): # lazily create causal attention mask, with full attention between the vision tokens # pytorch uses additive attention mask; fill with -inf mask = torch.empty(self.context_length, self.context_length) mask.fill_(float("-inf")) mask.triu_(1) # zero out the lower diagonal return mask @property def dtype(self): return self.visual.conv1.weight.dtype def encode_image(self, image, masks=None, pool_mask=None, dense=False): if pool_mask is not None: return self.visual(image.type(self.dtype), mask=pool_mask, dense=dense) if masks == None: return self.visual(image.type(self.dtype), dense=dense) else: return self.visual(image.type(self.dtype), masks.type(self.dtype)) def encode_text(self, text): x = self.token_embedding(text).type(self.dtype) # [batch_size, n_ctx, d_model] x = x + self.positional_embedding.type(self.dtype) x = x.permute(1, 0, 2) # NLD -> LND x = self.transformer(x) x = x.permute(1, 0, 2) # LND -> NLD x = self.ln_final(x).type(self.dtype) # x.shape = [batch_size, n_ctx, transformer.width] # take features from the eot embedding (eot_token is the highest number in each sequence) x = x[torch.arange(x.shape[0]), text.argmax(dim=-1)] @ self.text_projection return x def forward(self, image, text): image_features = self.encode_image(image) text_features = self.encode_text(text) # import pdb; pdb.set_trace() # normalized features # image_features shape: [1, 1024] image_features = image_features / image_features.norm(dim=-1, keepdim=True) text_features = text_features / text_features.norm(dim=-1, keepdim=True) # cosine similarity as logits logit_scale = self.logit_scale.exp() logits_per_iamge = logit_scale * image_features @ text_features.t() logits_per_text = logit_scale * text_features @ image_features.t() # shape = [global_batch_size, global_batch_size] return logits_per_iamge, logits_per_text def convert_weights(model: nn.Module): """Convert applicable model parameters to fp16""" def _convert_weights_to_fp16(l): if isinstance(l, (nn.Conv1d, nn.Conv2d, nn.Linear)): l.weight.data = l.weight.data.half() if l.bias is not None: l.bias.data = l.bias.data.half() if isinstance(l, nn.MultiheadAttention): for attr in [*[f"{s}_proj_weight" for s in ["in", "q", "k", "v"]], "in_proj_bias", "bias_k", "bias_v"]: tensor = getattr(l, attr) if tensor is not None: tensor.data = tensor.data.half() for name in ["text_projection", "proj"]: if hasattr(l, name): attr = getattr(l, name) if attr is not None: attr.data = attr.data.half() model.apply(_convert_weights_to_fp16) def build_model(state_dict: dict, prompt_depth=0, prompt_length=0): vit = "visual.proj" in state_dict if vit: vision_width = state_dict["visual.conv1.weight"].shape[0] vision_layers = len([k for k in state_dict.keys() if k.startswith("visual.") and k.endswith(".attn.in_proj_weight")]) vision_patch_size = state_dict["visual.conv1.weight"].shape[-1] grid_size = round((state_dict["visual.positional_embedding"].shape[0] - 1) ** 0.5) image_resolution = vision_patch_size * grid_size else: counts: list = [len(set(k.split(".")[2] for k in state_dict if k.startswith(f"visual.layer{b}"))) for b in [1, 2, 3, 4]] vision_layers = tuple(counts) vision_width = state_dict["visual.layer1.0.conv1.weight"].shape[0] output_width = round((state_dict["visual.attnpool.positional_embedding"].shape[0] - 1) ** 0.5) vision_patch_size = None assert output_width ** 2 + 1 == state_dict["visual.attnpool.positional_embedding"].shape[0] image_resolution = output_width * 32 embed_dim = state_dict["text_projection"].shape[1] context_length = state_dict["positional_embedding"].shape[0] vocab_size = state_dict["token_embedding.weight"].shape[0] transformer_width = state_dict["ln_final.weight"].shape[0] transformer_heads = transformer_width // 64 transformer_layers = len(set(k.split(".")[2] for k in state_dict if k.startswith(f"transformer.resblocks"))) model = CLIP( embed_dim, image_resolution, vision_layers, vision_width, vision_patch_size, context_length, vocab_size, transformer_width, transformer_heads, transformer_layers, prompt_depth=prompt_depth, prompt_length=prompt_length, ) for key in ["input_resolution", "context_length", "vocab_size"]: del state_dict[key] convert_weights(model) model.load_state_dict(state_dict, strict=False) return model.eval()