-
Notifications
You must be signed in to change notification settings - Fork 0
/
Mytrain.py
159 lines (142 loc) · 7.05 KB
/
Mytrain.py
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
import numpy as np
import matplotlib.pyplot as plt
import os
join = os.path.join
from tqdm import tqdm
import torch
from torch.utils.data import Dataset, DataLoader
import monai
import torch.nn as nn
from segment_anything import sam_model_registry
from segment_anything.utils.transforms import ResizeLongestSide
from segment_anything.modeling.CWDLoss import CriterionCWD
from torch.nn import functional as F
from torchvision.models.mobilenetv2 import InvertedResidual
# set seeds
torch.manual_seed(2024)
np.random.seed(2024)
def _upsample_like_1024(src):
src = F.interpolate(src, size=(1024, 1024), mode='bilinear')
return src
class NpzDataset(Dataset):
def __init__(self, data_root):
self.data_root = data_root
self.npz_files = sorted(os.listdir(self.data_root))
self.npz_data = [np.load(join(data_root, f)) for f in self.npz_files]
# this implementation is ugly but it works (and is also fast for feeding data to GPU)
# if your server has enough RAM
# as an alternative, you can also use a list of npy files and load them one by one
self.ori_gts = np.vstack([d['gts'] for d in self.npz_data])
self.ori_imgs = np.vstack([d['imgs'] for d in self.npz_data])
self.img_embeddings = np.vstack([d['img_embeddings'] for d in self.npz_data])
self.boundary = np.vstack([d['boundary'] for d in self.npz_data])
self.depth_embeddings = np.vstack([d['depth_embeddings'] for d in self.npz_data])
print(f"img_embeddings.shape={self.img_embeddings.shape}, ori_gts.shape={self.ori_gts.shape}, "
f"boundary.shape={self.boundary.shape}", f"depth_embeddings.shape={self.depth_embeddings.shape}")
def __len__(self):
return self.ori_gts.shape[0]
def __getitem__(self, index):
img_embed = self.img_embeddings[index]
gt2D = self.ori_gts[index]
img = self.ori_imgs[index]
boundary = self.boundary[index]
depth_embed = self.depth_embeddings[index]
y_indices, x_indices = np.where(gt2D > 0)
x_min, x_max = np.min(x_indices), np.max(x_indices)
y_min, y_max = np.min(y_indices), np.max(y_indices)
# add perturbation to bounding box coordinates
H, W = gt2D.shape
x_min = max(0, x_min - np.random.randint(0, 20))
x_max = min(W, x_max + np.random.randint(0, 20))
y_min = max(0, y_min - np.random.randint(0, 20))
y_max = min(H, y_max + np.random.randint(0, 20))
bboxes = np.array([x_min, y_min, x_max, y_max])
# convert img embedding, mask, bounding box to torch tensor
return torch.tensor(img_embed).float(), torch.tensor(img).float(), torch.tensor(gt2D[None, :, :]).long(), torch.tensor(bboxes).float(),\
torch.tensor(boundary[None, :, :]).long(), torch.tensor(depth_embed).float()
# %% test dataset class and dataloader
npz_tr_path = 'data/vit_b/COD_train'
work_dir = './work_dir_cod'
task_name = 'DSAM'
# prepare SAM model
model_type = 'vit_b'
checkpoint = 'work_dir_cod/SAM/sam_vit_b_01ec64.pth'
device = 'cuda:0'
model_save_path = join(work_dir, task_name)
os.makedirs(model_save_path, exist_ok=True)
sam_model = sam_model_registry[model_type](checkpoint=checkpoint).to(device)
sam_model.train()
# Set up the optimizer, hyperparameter tuning will improve performance here
optimizer = torch.optim.Adam(sam_model.mask_decoder.parameters(), lr=1e-5, weight_decay=0)
seg_loss = monai.losses.DiceCELoss(sigmoid=True, squared_pred=True, reduction='mean')
CWD_loss = CriterionCWD(norm_type='channel', divergence='kl', temperature=4.0)
num_epochs = 100
losses = []
best_loss = 1e10
train_dataset = NpzDataset(npz_tr_path)
mask_threshold = 0.0
train_dataloader = DataLoader(train_dataset, batch_size=8, shuffle=True)
for epoch in range(num_epochs+1):
epoch_loss = 0
# train
for step, (image_embedding, img, gt2D, boxes, boundary, depth_embedding) in enumerate(tqdm(train_dataloader)):
with torch.no_grad():
box_np = boxes.numpy()
sam_trans = ResizeLongestSide(sam_model.image_encoder.img_size)
box = sam_trans.apply_boxes(box_np, (gt2D.shape[-2], gt2D.shape[-1]))
box_torch = torch.as_tensor(box, dtype=torch.float, device=device)
boundary = torch.as_tensor(boundary, dtype=torch.float, device=device)
# boundary = boun_conv(boundary)
image_embedding = torch.as_tensor(image_embedding, dtype=torch.float, device=device)
depth_embedding = torch.as_tensor(depth_embedding, dtype=torch.float, device=device)
if len(box_torch.shape) == 2:
box_torch = box_torch[:, None, :] # (B, 1, 4)
# get prompt embeddings
sparse_embeddings_box, dense_embeddings_box = sam_model.prompt_encoder(
points=None,
boxes=box_torch,
masks=None
)
resize_img_tensor = np.transpose(img, (0, 3, 1, 2)).to(device)
input_image = _upsample_like_1024(resize_img_tensor)
pvt_embedding = sam_model.pvt(input_image)[3]
bc_embedding, pvt_64 = sam_model.BC(pvt_embedding)
# bc_embedding shape:" 1, 256, 64, 64
distill_loss = CWD_loss(bc_embedding, depth_embedding)
hybrid_embedding = torch.cat([pvt_64, bc_embedding], dim=1)
high_frequency = sam_model.DWT(hybrid_embedding)
dense_embeddings, sparse_embeddings = sam_model.ME(dense_embeddings_box,
high_frequency, sparse_embeddings_box)
# predicted masks
mask_predictions, _ = sam_model.mask_decoder(
image_embeddings=image_embedding.to(device), # (B, 256, 64, 64)
image_pe=sam_model.prompt_encoder.get_dense_pe(), # (1, 256, 64, 64)
sparse_prompt_embeddings=sparse_embeddings, # (B, 2, 256)
dense_prompt_embeddings=dense_embeddings, # (B, 256, 64, 64)
multimask_output=False,
)
final_mask = sam_model.loop_finer(mask_predictions, depth_embedding, depth_embedding)
mask_predictions = 0.1*final_mask + 0.9*mask_predictions
loss = 0.9*seg_loss(mask_predictions, gt2D.to(device)) + 0.1*distill_loss
optimizer.zero_grad()
loss.backward()
optimizer.step()
epoch_loss += loss.item()
epoch_loss /= step
losses.append(epoch_loss)
print(f'EPOCH: {epoch}, Loss: {epoch_loss}')
# save the latest model checkpoint
if epoch >= 80 and epoch % 10 == 0:
torch.save(sam_model.state_dict(), join(model_save_path, str(epoch) + 'sam_model.pth'))
# save the best model
if epoch_loss < best_loss:
best_loss = epoch_loss
torch.save(sam_model.state_dict(), join(model_save_path, 'sam_model_best.pth'))
# plot loss
plt.plot(losses)
plt.title('Dice + Cross Entropy Loss')
plt.xlabel('Epoch')
plt.ylabel('Loss')
# plt.show() # comment this line if you are running on a server
plt.savefig(join(model_save_path, 'train_loss.png'))
plt.close()