File size: 13,564 Bytes
e7610f7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
import torch
import matplotlib.pyplot as plt
from torchvision.utils import save_image
from torchvision import transforms
from torch.utils.data import DataLoader
import numpy as np, os
from torch import nn
import math
import torch.nn.functional as F
from torch.optim import Adam
from typing import Optional
import random


def mkdir(dir):
    if not os.path.exists(dir):
        os.makedirs(dir)


def get_beta_schedule(beta_schedule, beta_start, beta_end, num_diffusion_timesteps):
    def sigmoid(x):
        return 1 / (np.exp(-x) + 1)

    if beta_schedule == "quad":
        betas = (
            np.linspace(
                beta_start ** 0.5,
                beta_end ** 0.5,
                num_diffusion_timesteps,
                dtype=np.float64,
            )
            ** 2
        )

    elif beta_schedule == "linear":
        betas = np.linspace(
            beta_start, beta_end, num_diffusion_timesteps, dtype=np.float64
        )
    elif beta_schedule == "const":
        betas = beta_end * np.ones(num_diffusion_timesteps, dtype=np.float64)
    elif beta_schedule == "jsd":  # 1/T, 1/(T-1), 1/(T-2), ..., 1
        betas = 1.0 / np.linspace(
            num_diffusion_timesteps, 1, num_diffusion_timesteps, dtype=np.float64
        )
    elif beta_schedule == "sigmoid":
        betas = np.linspace(-6, 6, num_diffusion_timesteps)
        betas = sigmoid(betas) * (beta_end - beta_start) + beta_start
    else:
        raise NotImplementedError(beta_schedule)
    assert betas.shape == (num_diffusion_timesteps,)
    betas = torch.from_numpy(betas).float()
    return betas


def get_index_from_list(vals, t, x_shape):
    """ 
    Returns a specific index t of a passed list of values vals
    while considering the batch dimension.
    """
    batch_size = t.shape[0]
    out = vals.gather(-1, t.cpu())
    return out.reshape(batch_size, *((1,) * (len(x_shape) - 1))).to(t.device)


def forward_diffusion_sample(x, t, device="cpu"):
    """
    Takes an image and a timestep as input and 
    returns the noisy version of it
    """
    noise = torch.randn_like(x)  # gaussian noise
    # noise = torch.FloatTensor(x.shape).uniform_(-1, 1) #uniform distribution noise
    sqrt_alphas_cumprod_t = get_index_from_list(sqrt_alphas_cumprod, t, x.shape)
    sqrt_one_minus_alphas_cumprod_t = get_index_from_list(
        sqrt_one_minus_alphas_cumprod, t, x.shape
    )
    # print("coeff stats ",sqrt_alphas_cumprod_t, " and ", sqrt_one_minus_alphas_cumprod_t)
    # mean + variance
    return sqrt_alphas_cumprod_t.to(device) * x.to(device) \
           + sqrt_one_minus_alphas_cumprod_t.to(device) * noise.to(device), noise.to(device)


class Block(nn.Module):
    def __init__(self, in_ch, out_ch, time_emb_dim, up=False):
        super().__init__()
        self.time_mlp = nn.Linear(time_emb_dim, out_ch)
        if up:
            self.conv1 = nn.Conv2d(2 * in_ch, out_ch, 3, padding=1)
            self.transform = nn.ConvTranspose2d(out_ch, out_ch, 4, 2, 1)
        else:
            self.conv1 = nn.Conv2d(in_ch, out_ch, 3, padding=1)
            self.transform = nn.Conv2d(out_ch, out_ch, 4, 2, 1)
        self.conv2 = nn.Conv2d(out_ch, out_ch, 3, padding=1)
        self.bnorm1 = nn.BatchNorm2d(out_ch)
        self.bnorm2 = nn.BatchNorm2d(out_ch)
        self.relu = nn.LeakyReLU(0.2)

    def forward(self, x, t, ):
        # First Conv
        h = self.bnorm1(self.relu(self.conv1(x)))
        # Time embedding
        time_emb = self.relu(self.time_mlp(t))
        # Extend last 2 dimensions
        time_emb = time_emb[(...,) + (None,) * 2]
        # Add time channel
        h = h + time_emb
        # Second Conv
        h = self.bnorm2(self.relu(self.conv2(h)))
        # Down or Upsample
        return self.transform(h)


class SinusoidalPositionEmbeddings(nn.Module):
    def __init__(self, dim):
        super().__init__()
        self.dim = dim

    def forward(self, time):
        device = time.device
        half_dim = self.dim // 2
        embeddings = math.log(10000) / (half_dim - 1)
        embeddings = torch.exp(torch.arange(half_dim, device=device) * -embeddings)
        embeddings = time[:, None] * embeddings[None, :]
        embeddings = torch.cat((embeddings.sin(), embeddings.cos()), dim=-1)
        return embeddings


class CrossAttention(nn.Module):
    """
    ### Cross Attention Layer
    This falls-back to self-attention when conditional embeddings are not specified.
    """

    use_flash_attention: bool = True

    def __init__(self, d_model: int, d_cond: int, n_heads: int, d_head: int, is_inplace: bool = False):
        """
        :param d_model: is the input embedding size
        :param n_heads: is the number of attention heads
        :param d_head: is the size of a attention head
        :param d_cond: is the size of the conditional embeddings
        :param is_inplace: specifies whether to perform the attention softmax computation inplace to
            save memory
        """
        super().__init__()

        self.is_inplace = is_inplace
        self.n_heads = n_heads
        self.d_head = d_head

        # Attention scaling factor
        self.scale = d_head ** -0.5

        # Query, key and value mappings
        d_attn = d_head * n_heads
        self.to_q = nn.Linear(d_model, d_attn, bias=False)
        self.to_k = nn.Linear(d_cond, d_attn, bias=False)
        self.to_v = nn.Linear(d_cond, d_attn, bias=False)

        # Final linear layer
        self.to_out = nn.Sequential(nn.Linear(d_attn, d_model))

    def forward(self, x: torch.Tensor, cond: Optional[torch.Tensor] = None):
        """
        :param x: are the input embeddings of shape `[batch_size, height * width, d_model]`
        :param cond: is the conditional embeddings of shape `[batch_size, n_cond, d_cond]`
        """

        # If `cond` is `None` we perform self attention
        has_cond = cond is not None
        if not has_cond:
            cond = x

        # Get query, key and value vectors
        q = self.to_q(x)
        k = self.to_k(cond)
        v = self.to_v(cond)

        return self.normal_attention(q, k, v)

    def normal_attention(self, q: torch.Tensor, k: torch.Tensor, v: torch.Tensor):
        """
        #### Normal Attention

        :param q: are the query vectors before splitting heads, of shape `[batch_size, seq, d_attn]`
        :param k: are the query vectors before splitting heads, of shape `[batch_size, seq, d_attn]`
        :param v: are the query vectors before splitting heads, of shape `[batch_size, seq, d_attn]`
        """

        # Split them to heads of shape `[batch_size, seq_len, n_heads, d_head]`
        q = q.view(*q.shape[:2], self.n_heads, -1)
        k = k.view(*k.shape[:2], self.n_heads, -1)
        v = v.view(*v.shape[:2], self.n_heads, -1)

        # Calculate attention $\frac{Q K^\top}{\sqrt{d_{key}}}$
        attn = torch.einsum('bihd,bjhd->bhij', q, k) * self.scale

        # Compute softmax
        # $$\underset{seq}{softmax}\Bigg(\frac{Q K^\top}{\sqrt{d_{key}}}\Bigg)$$
        if self.is_inplace:
            half = attn.shape[0] // 2
            attn[half:] = attn[half:].softmax(dim=-1)
            attn[:half] = attn[:half].softmax(dim=-1)
        else:
            attn = attn.softmax(dim=-1)

        # Compute attention output
        # $$\underset{seq}{softmax}\Bigg(\frac{Q K^\top}{\sqrt{d_{key}}}\Bigg)V$$
        out = torch.einsum('bhij,bjhd->bihd', attn, v)
        # Reshape to `[batch_size, height * width, n_heads * d_head]`
        out = out.reshape(*out.shape[:2], -1)
        # Map to `[batch_size, height * width, d_model]` with a linear layer
        return self.to_out(out)


class SimpleUnet(nn.Module):
    def __init__(self):
        super().__init__()
        image_channels = 3
        # down_channels = (64, 128, 256, 512, 1024)
        # up_channels = (1024, 512, 256, 128, 64)
        down_channels = (16, 32, 64, 128, 256)
        up_channels = (256, 128, 64, 32, 16)
        out_dim = 1
        time_emb_dim = 32

        # Time embedding
        self.time_mlp = nn.Sequential(
            SinusoidalPositionEmbeddings(time_emb_dim),
            nn.Linear(time_emb_dim, time_emb_dim),
            nn.ReLU()
        )

        # Initial projection
        self.conv0 = nn.Conv2d(image_channels, down_channels[0], 3, padding=1)

        # Downsample
        self.downs = nn.ModuleList([Block(down_channels[i], down_channels[i + 1], \
                                          time_emb_dim) \
                                    for i in range(len(down_channels) - 1)])
        # Upsample
        self.ups = nn.ModuleList([Block(up_channels[i], up_channels[i + 1], \
                                        time_emb_dim, up=True) \
                                  for i in range(len(up_channels) - 1)])

        self.silu = nn.SiLU()

        self.output = nn.Conv2d(up_channels[-1], 3, out_dim)

        self.apply_tanh = nn.Tanh()

        self.cross_attention_module = CrossAttention(3, 32, 16, 16)

    def forward(self, x, y, timestep):

        # Embedd class condition using cross attention
        batch_size = x.shape[0]
        y = self.time_mlp(y)
        y = y[:, None, :]
        x = x.permute(0, 2, 3, 1).view(batch_size, IMG_SIZE * IMG_SIZE, 3)
        x2 = x + self.cross_attention_module(x, y)
        x2 = x2.view(batch_size, IMG_SIZE, IMG_SIZE, 3).permute(0, 3, 1, 2)

        # Embedd time
        t = self.time_mlp(timestep)

        # Initial conv
        x2 = self.conv0(x2)

        # Unet
        residual_inputs = []
        for down in self.downs:
            x2 = down(x2, t)
            residual_inputs.append(x2)
        for up in self.ups:
            residual_x2 = residual_inputs.pop()
            # Add residual x2 as additional channels
            x2 = torch.cat((x2, residual_x2), dim=1)
            x2 = up(x2, t)

        x2 = self.silu(x2)

        output = self.output(x2)

        return output


def get_loss(model, x_0, t):
    latent, condition = x_0  # both latents and condition have same shap
    latent = latent.cuda()
    condition = condition.cuda()
    x_noisy, noise = forward_diffusion_sample(latent, t, device)
    noise_pred = model(x_noisy, condition, t)

    # return F.l1_loss(noise, noise_pred)
    return F.mse_loss(noise, noise_pred)


@torch.no_grad()
def sample_timestep(x, model, y, t):
    betas_t = get_index_from_list(betas, t, x.shape)
    sqrt_one_minus_alphas_cumprod_t = get_index_from_list(
        sqrt_one_minus_alphas_cumprod, t, x.shape
    )
    sqrt_recip_alphas_t = get_index_from_list(sqrt_recip_alphas, t, x.shape)

    # Call model (current image - noise prediction)
    model_mean = sqrt_recip_alphas_t * (
        x - (betas_t / sqrt_one_minus_alphas_cumprod_t) * model(x, y, t)
    )
    posterior_variance_t = get_index_from_list(posterior_variance, t, x.shape)

    # print("model prediction stats ",torch.max(model(x, y, t)), " and ", torch.min(model(x, y, t)))

    if t == 0:
        return model_mean
    else:
        noise = torch.randn_like(x)
        return model_mean + torch.sqrt(posterior_variance_t) * noise


def show_tensor_image(image):
    reverse_transforms = transforms.Compose([
        transforms.Lambda(lambda t: (t + 1) / 2),
        transforms.Lambda(lambda t: t.permute(1, 2, 0)),  # CHW to HWC
        transforms.Lambda(lambda t: t * 255.),
        transforms.Lambda(lambda t: t.numpy().astype(np.uint8)),
        transforms.ToPILImage(),
    ])

    # Take first image of batch
    if len(image.shape) == 4:
        image = image[0, :, :, :]
    plt.imshow(reverse_transforms(image))


def generate_latent(model_dir, cancer_type, output_dir):
    if (cancer_type == 'benign'):
        model_name = "digestpath_mask_benign_default.pt"
    else:
        model_name = "digestpath_mask_malignant_default.pt"

    device = "cuda" if torch.cuda.is_available() else "cpu"


    model_path = os.path.join(model_dir, model_name)

    model = SimpleUnet()
    model.to(device)

    model.load_state_dict(torch.load(model_path))
    print("model loaded")
    model.eval()

    # cancer_grade = random.randint(0, 1)
    condition = torch.tensor([1]).cuda()  # benign:0/malignant:1 grade cancer
    # condition = torch.full([1, 1, IMG_SIZE, IMG_SIZE], condition).float().cuda()
    img = torch.randn((1, 3, IMG_SIZE, IMG_SIZE), device=device)
    for j in range(0, T)[::-1]:
        t = torch.full((1,), j, device=device, dtype=torch.long)
        img = sample_timestep(img, model, condition, t)
    print("sampled image ", torch.max(img), " and ", torch.min(img))
    save_image(img, os.path.join(output_dir, "sample.png"))
    torch.save(img, os.path.join(output_dir, "sample.pt"))

# Define beta schedule
T = 1000
IMG_SIZE = 64

betas = get_beta_schedule(beta_schedule="linear",
                          beta_start=0.0001,
                          beta_end=0.02,
                          num_diffusion_timesteps=T)

# Pre-calculate different terms for closed form
alphas = 1. - betas
alphas_cumprod = torch.cumprod(alphas, axis=0)
sqrt_alphas_cumprod = torch.sqrt(alphas_cumprod)  # root(alpha_bar)
sqrt_one_minus_alphas_cumprod = torch.sqrt(1. - alphas_cumprod)  # root(1-alpha_bar)
sqrt_recip_alphas = torch.sqrt(1.0 / alphas)  # 1/root(alpha)
alphas_cumprod_prev = F.pad(alphas_cumprod[:-1], (1, 0), value=1.0)
posterior_variance = betas * (1. - alphas_cumprod_prev) / (1. - alphas_cumprod)


# model_dir = "trained_models/diffusion"
# output_dir = r"F:\Datasets\DigestPath\scene_generation\all\1000\256\test\output\benign"
# generate_latent(model_dir, 'malignant', output_dir)