|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
""" Nested Transformer (NesT) in PyTorch |
|
A PyTorch implement of Aggregating Nested Transformers as described in: |
|
'Aggregating Nested Transformers' |
|
- https://arxiv.org/abs/2105.12723 |
|
The official Jax code is released and available at https://github.com/google-research/nested-transformer. |
|
The weights have been converted with convert/convert_nest_flax.py |
|
Acknowledgments: |
|
* The paper authors for sharing their research, code, and model weights |
|
* Ross Wightman's existing code off which I based this |
|
Copyright 2021 Alexander Soare |
|
|
|
""" |
|
|
|
import collections.abc |
|
import logging |
|
import math |
|
from functools import partial |
|
from typing import Callable, Sequence |
|
|
|
import torch |
|
import torch.nn.functional as F |
|
from torch import nn |
|
|
|
from .nest import DropPath, Mlp, _assert, create_conv3d, create_pool3d, to_ntuple, trunc_normal_ |
|
from .patchEmbed3D import PatchEmbed3D |
|
|
|
_logger = logging.getLogger(__name__) |
|
|
|
|
|
class Attention(nn.Module): |
|
""" |
|
This is much like `.vision_transformer.Attention` but uses *localised* self attention by accepting an input with |
|
an extra "image block" dim |
|
""" |
|
|
|
def __init__(self, dim, num_heads=8, qkv_bias=False, attn_drop=0.0, proj_drop=0.0): |
|
super().__init__() |
|
self.num_heads = num_heads |
|
head_dim = dim // num_heads |
|
self.scale = head_dim**-0.5 |
|
|
|
self.qkv = nn.Linear(dim, 3 * dim, bias=qkv_bias) |
|
self.attn_drop = nn.Dropout(attn_drop) |
|
self.proj = nn.Linear(dim, dim) |
|
self.proj_drop = nn.Dropout(proj_drop) |
|
|
|
def forward(self, x): |
|
""" |
|
x is shape: B (batch_size), T (image blocks), N (seq length per image block), C (embed dim) |
|
""" |
|
b, t, n, c = x.shape |
|
|
|
qkv = self.qkv(x).reshape(b, t, n, 3, self.num_heads, c // self.num_heads).permute(3, 0, 4, 1, 2, 5) |
|
q, k, v = qkv.unbind(0) |
|
|
|
attn = (q @ k.transpose(-2, -1)) * self.scale |
|
attn = attn.softmax(dim=-1) |
|
attn = self.attn_drop(attn) |
|
|
|
x = (attn @ v).permute(0, 2, 3, 4, 1).reshape(b, t, n, c) |
|
x = self.proj(x) |
|
x = self.proj_drop(x) |
|
return x |
|
|
|
|
|
class TransformerLayer(nn.Module): |
|
""" |
|
This is much like `.vision_transformer.Block` but: |
|
- Called TransformerLayer here to allow for "block" as defined in the paper ("non-overlapping image blocks") |
|
- Uses modified Attention layer that handles the "block" dimension |
|
""" |
|
|
|
def __init__( |
|
self, |
|
dim, |
|
num_heads, |
|
mlp_ratio=4.0, |
|
qkv_bias=False, |
|
drop=0.0, |
|
attn_drop=0.0, |
|
drop_path=0.0, |
|
act_layer=nn.GELU, |
|
norm_layer=nn.LayerNorm, |
|
): |
|
super().__init__() |
|
self.norm1 = norm_layer(dim) |
|
self.attn = Attention(dim, num_heads=num_heads, qkv_bias=qkv_bias, attn_drop=attn_drop, proj_drop=drop) |
|
self.drop_path = DropPath(drop_path) if drop_path > 0.0 else nn.Identity() |
|
self.norm2 = norm_layer(dim) |
|
mlp_hidden_dim = int(dim * mlp_ratio) |
|
self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop) |
|
|
|
def forward(self, x): |
|
y = self.norm1(x) |
|
x = x + self.drop_path(self.attn(y)) |
|
x = x + self.drop_path(self.mlp(self.norm2(x))) |
|
return x |
|
|
|
|
|
class ConvPool(nn.Module): |
|
def __init__(self, in_channels, out_channels, norm_layer, pad_type=""): |
|
super().__init__() |
|
self.conv = create_conv3d(in_channels, out_channels, kernel_size=3, padding=pad_type, bias=True) |
|
self.norm = norm_layer(out_channels) |
|
self.pool = create_pool3d("max", kernel_size=3, stride=2, padding=pad_type) |
|
|
|
def forward(self, x): |
|
""" |
|
x is expected to have shape (B, C, D, H, W) |
|
""" |
|
_assert(x.shape[-3] % 2 == 0, "BlockAggregation requires even input spatial dims") |
|
_assert(x.shape[-2] % 2 == 0, "BlockAggregation requires even input spatial dims") |
|
_assert(x.shape[-1] % 2 == 0, "BlockAggregation requires even input spatial dims") |
|
|
|
|
|
x = self.conv(x) |
|
|
|
x = self.norm(x.permute(0, 2, 3, 4, 1)).permute(0, 4, 1, 2, 3) |
|
x = self.pool(x) |
|
return x |
|
|
|
|
|
def blockify(x, block_size: int): |
|
"""image to blocks |
|
Args: |
|
x (Tensor): with shape (B, D, H, W, C) |
|
block_size (int): edge length of a single square block in units of D, H, W |
|
""" |
|
b, d, h, w, c = x.shape |
|
_assert(d % block_size == 0, "`block_size` must divide input depth evenly") |
|
_assert(h % block_size == 0, "`block_size` must divide input height evenly") |
|
_assert(w % block_size == 0, "`block_size` must divide input width evenly") |
|
grid_depth = d // block_size |
|
grid_height = h // block_size |
|
grid_width = w // block_size |
|
x = x.reshape(b, grid_depth, block_size, grid_height, block_size, grid_width, block_size, c) |
|
|
|
x = x.permute(0, 1, 3, 5, 2, 4, 6, 7).reshape( |
|
b, grid_depth * grid_height * grid_width, -1, c |
|
) |
|
|
|
return x |
|
|
|
|
|
|
|
def deblockify(x, block_size: int): |
|
"""blocks to image |
|
Args: |
|
x (Tensor): with shape (B, T, N, C) where T is number of blocks and N is sequence size per block |
|
block_size (int): edge length of a single square block in units of desired D, H, W |
|
""" |
|
b, t, _, c = x.shape |
|
grid_size = round(math.pow(t, 1 / 3)) |
|
depth = height = width = grid_size * block_size |
|
x = x.reshape(b, grid_size, grid_size, grid_size, block_size, block_size, block_size, c) |
|
|
|
x = x.permute(0, 1, 4, 2, 5, 3, 6, 7).reshape(b, depth, height, width, c) |
|
|
|
return x |
|
|
|
|
|
class NestLevel(nn.Module): |
|
"""Single hierarchical level of a Nested Transformer""" |
|
|
|
def __init__( |
|
self, |
|
num_blocks, |
|
block_size, |
|
seq_length, |
|
num_heads, |
|
depth, |
|
embed_dim, |
|
prev_embed_dim=None, |
|
mlp_ratio=4.0, |
|
qkv_bias=True, |
|
drop_rate=0.0, |
|
attn_drop_rate=0.0, |
|
drop_path_rates: Sequence[int] = (), |
|
norm_layer=None, |
|
act_layer=None, |
|
pad_type="", |
|
): |
|
super().__init__() |
|
self.block_size = block_size |
|
self.pos_embed = nn.Parameter(torch.zeros(1, num_blocks, seq_length, embed_dim)) |
|
|
|
if prev_embed_dim is not None: |
|
self.pool = ConvPool(prev_embed_dim, embed_dim, norm_layer=norm_layer, pad_type=pad_type) |
|
else: |
|
self.pool = nn.Identity() |
|
|
|
|
|
if len(drop_path_rates): |
|
assert len(drop_path_rates) == depth, "Must provide as many drop path rates as there are transformer layers" |
|
self.transformer_encoder = nn.Sequential( |
|
*[ |
|
TransformerLayer( |
|
dim=embed_dim, |
|
num_heads=num_heads, |
|
mlp_ratio=mlp_ratio, |
|
qkv_bias=qkv_bias, |
|
drop=drop_rate, |
|
attn_drop=attn_drop_rate, |
|
drop_path=drop_path_rates[i], |
|
norm_layer=norm_layer, |
|
act_layer=act_layer, |
|
) |
|
for i in range(depth) |
|
] |
|
) |
|
|
|
def forward(self, x): |
|
""" |
|
expects x as (B, C, D, H, W) |
|
""" |
|
x = self.pool(x) |
|
x = x.permute(0, 2, 3, 4, 1) |
|
|
|
x = blockify(x, self.block_size) |
|
x = x + self.pos_embed |
|
|
|
x = self.transformer_encoder(x) |
|
|
|
x = deblockify(x, self.block_size) |
|
|
|
return x.permute(0, 4, 1, 2, 3) |
|
|
|
|
|
class NestTransformer3D(nn.Module): |
|
"""Nested Transformer (NesT) |
|
A PyTorch impl of : `Aggregating Nested Transformers` |
|
- https://arxiv.org/abs/2105.12723 |
|
""" |
|
|
|
def __init__( |
|
self, |
|
img_size=96, |
|
in_chans=1, |
|
patch_size=2, |
|
num_levels=3, |
|
embed_dims=(128, 256, 512), |
|
num_heads=(4, 8, 16), |
|
depths=(2, 2, 20), |
|
num_classes=1000, |
|
mlp_ratio=4.0, |
|
qkv_bias=True, |
|
drop_rate=0.0, |
|
attn_drop_rate=0.0, |
|
drop_path_rate=0.5, |
|
norm_layer=None, |
|
act_layer=None, |
|
pad_type="", |
|
weight_init="", |
|
global_pool="avg", |
|
): |
|
""" |
|
Args: |
|
img_size (int, tuple): input image size |
|
in_chans (int): number of input channels |
|
patch_size (int): patch size |
|
num_levels (int): number of block hierarchies (T_d in the paper) |
|
embed_dims (int, tuple): embedding dimensions of each level |
|
num_heads (int, tuple): number of attention heads for each level |
|
depths (int, tuple): number of transformer layers for each level |
|
num_classes (int): number of classes for classification head |
|
mlp_ratio (int): ratio of mlp hidden dim to embedding dim for MLP of transformer layers |
|
qkv_bias (bool): enable bias for qkv if True |
|
drop_rate (float): dropout rate for MLP of transformer layers, MSA final projection layer, and classifier |
|
attn_drop_rate (float): attention dropout rate |
|
drop_path_rate (float): stochastic depth rate |
|
norm_layer: (nn.Module): normalization layer for transformer layers |
|
act_layer: (nn.Module): activation layer in MLP of transformer layers |
|
pad_type: str: Type of padding to use '' for PyTorch symmetric, 'same' for TF SAME |
|
weight_init: (str): weight init scheme |
|
global_pool: (str): type of pooling operation to apply to final feature map |
|
Notes: |
|
- Default values follow NesT-B from the original Jax code. |
|
- `embed_dims`, `num_heads`, `depths` should be ints or tuples with length `num_levels`. |
|
- For those following the paper, Table A1 may have errors! |
|
- https://github.com/google-research/nested-transformer/issues/2 |
|
""" |
|
super().__init__() |
|
|
|
for param_name in ["embed_dims", "num_heads", "depths"]: |
|
param_value = locals()[param_name] |
|
if isinstance(param_value, collections.abc.Sequence): |
|
assert len(param_value) == num_levels, f"Require `len({param_name}) == num_levels`" |
|
|
|
embed_dims = to_ntuple(num_levels)(embed_dims) |
|
num_heads = to_ntuple(num_levels)(num_heads) |
|
depths = to_ntuple(num_levels)(depths) |
|
self.num_classes = num_classes |
|
self.num_features = embed_dims[-1] |
|
self.feature_info = [] |
|
norm_layer = norm_layer or partial(nn.LayerNorm, eps=1e-6) |
|
act_layer = act_layer or nn.GELU |
|
self.drop_rate = drop_rate |
|
self.num_levels = num_levels |
|
if isinstance(img_size, collections.abc.Sequence): |
|
assert img_size[0] == img_size[1], "Model only handles square inputs" |
|
img_size = img_size[0] |
|
assert img_size % patch_size == 0, "`patch_size` must divide `img_size` evenly" |
|
self.patch_size = patch_size |
|
|
|
|
|
self.num_blocks = (8 ** torch.arange(num_levels)).flip(0).tolist() |
|
assert (img_size // patch_size) % round( |
|
math.pow(self.num_blocks[0], 1 / 3) |
|
) == 0, "First level blocks don't fit evenly. Check `img_size`, `patch_size`, and `num_levels`" |
|
|
|
|
|
|
|
|
|
self.block_size = int((img_size // patch_size) // round(math.pow(self.num_blocks[0], 1 / 3))) |
|
|
|
|
|
self.patch_embed = PatchEmbed3D( |
|
img_size=[img_size, img_size, img_size], |
|
patch_size=[patch_size, patch_size, patch_size], |
|
in_chans=in_chans, |
|
embed_dim=embed_dims[0], |
|
) |
|
self.num_patches = self.patch_embed.num_patches |
|
self.seq_length = self.num_patches // self.num_blocks[0] |
|
|
|
levels = [] |
|
|
|
dp_rates = [x.tolist() for x in torch.linspace(0, drop_path_rate, sum(depths)).split(depths)] |
|
prev_dim = None |
|
curr_stride = 4 |
|
for i in range(len(self.num_blocks)): |
|
dim = embed_dims[i] |
|
levels.append( |
|
NestLevel( |
|
self.num_blocks[i], |
|
self.block_size, |
|
self.seq_length, |
|
num_heads[i], |
|
depths[i], |
|
dim, |
|
prev_dim, |
|
mlp_ratio, |
|
qkv_bias, |
|
drop_rate, |
|
attn_drop_rate, |
|
dp_rates[i], |
|
norm_layer, |
|
act_layer, |
|
pad_type=pad_type, |
|
) |
|
) |
|
self.feature_info += [dict(num_chs=dim, reduction=curr_stride, module=f"levels.{i}")] |
|
prev_dim = dim |
|
curr_stride *= 2 |
|
|
|
self.levels = nn.ModuleList([levels[i] for i in range(num_levels)]) |
|
|
|
|
|
self.norm = norm_layer(embed_dims[-1]) |
|
|
|
self.init_weights(weight_init) |
|
|
|
def init_weights(self, mode=""): |
|
assert mode in ("nlhb", "") |
|
head_bias = -math.log(self.num_classes) if "nlhb" in mode else 0.0 |
|
for level in self.levels: |
|
trunc_normal_(level.pos_embed, std=0.02, a=-2, b=2) |
|
named_apply(partial(_init_nest_weights, head_bias=head_bias), self) |
|
|
|
@torch.jit.ignore |
|
def no_weight_decay(self): |
|
return {f"level.{i}.pos_embed" for i in range(len(self.levels))} |
|
|
|
def get_classifier(self): |
|
return self.head |
|
|
|
def forward_features(self, x): |
|
"""x shape (B, C, D, H, W)""" |
|
x = self.patch_embed(x) |
|
|
|
hidden_states_out = [x] |
|
|
|
for _, level in enumerate(self.levels): |
|
x = level(x) |
|
hidden_states_out.append(x) |
|
|
|
x = self.norm(x.permute(0, 2, 3, 4, 1)).permute(0, 4, 1, 2, 3) |
|
return x, hidden_states_out |
|
|
|
def forward(self, x): |
|
"""x shape (B, C, D, H, W)""" |
|
x = self.forward_features(x) |
|
|
|
if self.drop_rate > 0.0: |
|
x = F.dropout(x, p=self.drop_rate, training=self.training) |
|
return x |
|
|
|
|
|
def named_apply(fn: Callable, module: nn.Module, name="", depth_first=True, include_root=False) -> nn.Module: |
|
if not depth_first and include_root: |
|
fn(module=module, name=name) |
|
for child_name, child_module in module.named_children(): |
|
child_name = ".".join((name, child_name)) if name else child_name |
|
named_apply(fn=fn, module=child_module, name=child_name, depth_first=depth_first, include_root=True) |
|
if depth_first and include_root: |
|
fn(module=module, name=name) |
|
return module |
|
|
|
|
|
def _init_nest_weights(module: nn.Module, name: str = "", head_bias: float = 0.0): |
|
"""NesT weight initialization |
|
Can replicate Jax implementation. Otherwise follows vision_transformer.py |
|
""" |
|
if isinstance(module, nn.Linear): |
|
if name.startswith("head"): |
|
trunc_normal_(module.weight, std=0.02, a=-2, b=2) |
|
nn.init.constant_(module.bias, head_bias) |
|
else: |
|
trunc_normal_(module.weight, std=0.02, a=-2, b=2) |
|
if module.bias is not None: |
|
nn.init.zeros_(module.bias) |
|
elif isinstance(module, nn.Conv2d): |
|
trunc_normal_(module.weight, std=0.02, a=-2, b=2) |
|
if module.bias is not None: |
|
nn.init.zeros_(module.bias) |
|
elif isinstance(module, (nn.LayerNorm, nn.GroupNorm, nn.BatchNorm2d)): |
|
nn.init.zeros_(module.bias) |
|
nn.init.ones_(module.weight) |
|
|
|
|
|
def resize_pos_embed(posemb, posemb_new): |
|
""" |
|
Rescale the grid of position embeddings when loading from state_dict |
|
Expected shape of position embeddings is (1, T, N, C), and considers only square images |
|
""" |
|
_logger.info("Resized position embedding: %s to %s", posemb.shape, posemb_new.shape) |
|
seq_length_old = posemb.shape[2] |
|
num_blocks_new, seq_length_new = posemb_new.shape[1:3] |
|
size_new = int(math.sqrt(num_blocks_new * seq_length_new)) |
|
|
|
posemb = deblockify(posemb, int(math.sqrt(seq_length_old))).permute(0, 3, 1, 2) |
|
posemb = F.interpolate(posemb, size=[size_new, size_new], mode="bicubic", align_corners=False) |
|
|
|
posemb = blockify(posemb.permute(0, 2, 3, 1), int(math.sqrt(seq_length_new))) |
|
return posemb |
|
|
|
|
|
def checkpoint_filter_fn(state_dict, model): |
|
"""resize positional embeddings of pretrained weights""" |
|
pos_embed_keys = [k for k in state_dict.keys() if k.startswith("pos_embed_")] |
|
for k in pos_embed_keys: |
|
if state_dict[k].shape != getattr(model, k).shape: |
|
state_dict[k] = resize_pos_embed(state_dict[k], getattr(model, k)) |
|
return state_dict |
|
|