Hugging Face's logo

arjundd
/
noise2recon-release

English
mri
reconstruction
denoising
Model card Files Community
noise2recon-release
File size: 4,276 Bytes 
a226c5d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
AUG_TEST:
  UNDERSAMPLE:
    ACCELERATIONS:
    - 12
AUG_TRAIN:
  MOTION_P: 0.2
  MRI_RECON:
    AUG_SENSITIVITY_MAPS: true
    SCHEDULER_P:
      IGNORE: false
    TRANSFORMS: []
  NOISE_P: 0.2
  UNDERSAMPLE:
    ACCELERATIONS:
    - 12
    CALIBRATION_SIZE: 20
    CENTER_FRACTIONS: []
    MAX_ATTEMPTS: 30
    NAME: PoissonDiskMaskFunc
  USE_MOTION: false
  USE_NOISE: false
CUDNN_BENCHMARK: false
DATALOADER:
  ALT_SAMPLER:
    PERIOD_SUPERVISED: 1
    PERIOD_UNSUPERVISED: 1
  DATA_KEYS: []
  DROP_LAST: true
  FILTER:
    BY: []
  GROUP_SAMPLER:
    AS_BATCH_SAMPLER: false
    BATCH_BY: []
  NUM_WORKERS: 12
  PREFETCH_FACTOR: 2
  SAMPLER_TRAIN: ''
  SUBSAMPLE_TRAIN:
    NUM_TOTAL: 3
    NUM_TOTAL_BY_GROUP: []
    NUM_UNDERSAMPLED: 0
    NUM_VAL: -1
    NUM_VAL_BY_GROUP: []
    SEED: 1000
DATASETS:
  TEST:
  - mridata_knee_2019_test
  TRAIN:
  - mridata_knee_2019_train
  VAL:
  - mridata_knee_2019_val
DESCRIPTION:
  BRIEF: Denoiser finetuning U-Net 12x - std_dev=(0.2, 0.5), 3 supervised scan(s),
    0 unsupervised scan(s), 40k iterations, restart-iteration-count = True
  ENTITY_NAME: ''
  EXP_NAME: noise2recon/mridata_knee_3dfse/12x/Supervised-FT_3sub
  PROJECT_NAME: ''
  TAGS:
  - baseline
  - supervised-ft
  - 12x
  - mridata_knee_3dfse
MODEL:
  A2R:
    META_ARCHITECTURE: GeneralizedUnrolledCNN
    USE_SUPERVISED_CONSISTENCY: false
  CONSISTENCY:
    AUG:
      MOTION:
        RANGE:
        - 0.2
        - 0.5
        SCHEDULER:
          WARMUP_ITERS: 0
          WARMUP_METHOD: ''
      MRI_RECON:
        AUG_SENSITIVITY_MAPS: true
        SCHEDULER_P:
          IGNORE: false
        TRANSFORMS: []
      NOISE:
        MASK:
          RHO: 1.0
        SCHEDULER:
          WARMUP_ITERS: 0
          WARMUP_METHOD: ''
        STD_DEV:
        - 1
    LATENT_LOSS_NAME: mag_l1
    LATENT_LOSS_WEIGHT: 0.1
    LOSS_NAME: l1
    LOSS_WEIGHT: 0.1
    NUM_LATENT_LAYERS: 1
    USE_CONSISTENCY: true
    USE_LATENT: false
  CS:
    MAX_ITER: 200
    REGULARIZATION: 0.005
  DENOISING:
    META_ARCHITECTURE: GeneralizedUnrolledCNN
    NOISE:
      STD_DEV:
      - 0.2
      - 0.5
      USE_FULLY_SAMPLED_TARGET: true
      USE_FULLY_SAMPLED_TARGET_EVAL: null
  DEVICE: cpu
  M2R:
    META_ARCHITECTURE: GeneralizedUnrolledCNN
    USE_SUPERVISED_CONSISTENCY: false
  META_ARCHITECTURE: UnetModel
  N2R:
    META_ARCHITECTURE: GeneralizedUnrolledCNN
    USE_SUPERVISED_CONSISTENCY: false
  NM2R:
    META_ARCHITECTURE: GeneralizedUnrolledCNN
    USE_SUPERVISED_CONSISTENCY: false
  NORMALIZER:
    KEYWORDS: []
    NAME: TopMagnitudeNormalizer
  RECON_LOSS:
    NAME: l1
    RENORMALIZE_DATA: false
  SEG:
    ACTIVATION: sigmoid
    CLASSES: []
    INCLUDE_BACKGROUND: false
  SSDU:
    MASKER:
      PARAMS: {}
    META_ARCHITECTURE: GeneralizedUnrolledCNN
  UNET:
    BLOCK_ORDER:
    - conv
    - relu
    - conv
    - relu
    - batchnorm
    - dropout
    CHANNELS: 32
    DROPOUT: 0.0
    IN_CHANNELS: 2
    NORMALIZE: false
    NUM_POOL_LAYERS: 4
    OUT_CHANNELS: 2
  UNROLLED:
    BLOCK_ARCHITECTURE: ResNet
    CONV_BLOCK:
      ACTIVATION: relu
      NORM: none
      NORM_AFFINE: false
      ORDER:
      - norm
      - act
      - drop
      - conv
    DROPOUT: 0.0
    FIX_STEP_SIZE: false
    KERNEL_SIZE:
    - 3
    NUM_EMAPS: 1
    NUM_FEATURES: 256
    NUM_RESBLOCKS: 2
    NUM_UNROLLED_STEPS: 5
    PADDING: ''
    SHARE_WEIGHTS: false
  WEIGHTS: /bmrNAS/people/arjun/results/ss_recon/n2r-experiments/denoise-pretrainer/version_015/model_0036399.pth
OUTPUT_DIR: results://noise2recon/mridata_knee_3dfse/12x/Supervised-FT_3sub
SEED: 1000
SOLVER:
  BASE_LR: 0.001
  BIAS_LR_FACTOR: 1.0
  CHECKPOINT_PERIOD: -10
  GAMMA: 0.1
  GRAD_ACCUM_ITERS: 1
  LR_SCHEDULER_NAME: ''
  MAX_ITER: 40000
  MOMENTUM: 0.9
  OPTIMIZER: Adam
  STEPS:
  - 320000
  TEST_BATCH_SIZE: 16
  TRAIN_BATCH_SIZE: 16
  WARMUP_FACTOR: 0.001
  WARMUP_ITERS: 1000
  WARMUP_METHOD: linear
  WEIGHT_DECAY: 0.0001
  WEIGHT_DECAY_BIAS: 0.0001
  WEIGHT_DECAY_NORM: 0.0
TEST:
  EVAL_PERIOD: -10
  EXPECTED_RESULTS: []
  FLUSH_PERIOD: 0
  VAL_AS_TEST: true
  VAL_METRICS:
    RECON:
    - psnr
    - psnr_scan
    - psnr_mag
    - psnr_mag_scan
    - nrmse
    - nrmse_scan
    - nrmse_mag
    - nrmse_mag_scan
    - ssim (Wang)
TIME_SCALE: iter
VERSION: 1
VIS_PERIOD: 400