mmagic swinir训练去噪任务时配置文件问题

[wsllkkl]

配置文件如下：
base = [
'../base/default_runtime.py',
'../base/datasets/denoising-gaussian_gray_test_config.py'
]

experiment_name = 'swinir_s128w8d6e180_8xb1-lr2e-4-1600k_dfwb-grayDN15'
work_dir = f'./work_dirs/{experiment_name}'
save_dir = './work_dirs/'

modify sigma of RandomNoise

sigma = 15
test_dataloader = base.test_dataloader
for dataloader in test_dataloader:
test_pipeline = dataloader['dataset']['pipeline']
test_pipeline[2]['params']['gaussian_sigma'] = [sigma, sigma]

model settings

model = dict(
type='BaseEditModel',
generator=dict(
type='SwinIRNet',
upscale=1,
in_chans=1,
img_size=128,
window_size=8,
img_range=1.0,
depths=[6, 6, 6, 6, 6, 6],
embed_dim=180,
num_heads=[6, 6, 6, 6, 6, 6],
mlp_ratio=2,
upsampler='',
resi_connection='1conv'),
pixel_loss=dict(type='CharbonnierLoss', eps=1e-9),
data_preprocessor=dict(type='DataPreprocessor', mean=[0.], std=[255.]))

train_pipeline = [
dict(
type='LoadImageFromFile',
key='img',
color_type='grayscale',
imdecode_backend='cv2'),
dict(
type='LoadImageFromFile',
key='gt',
color_type='grayscale',
imdecode_backend='cv2'),
dict(type='SetValues', dictionary=dict(scale=1)),
dict(type='PairedRandomCrop', gt_patch_size=128),
dict(
type='Flip',
keys=['img', 'gt'],
flip_ratio=0.5,
direction='horizontal'),
dict(
type='Flip', keys=['img', 'gt'], flip_ratio=0.5, direction='vertical'),
dict(type='RandomTransposeHW', keys=['img', 'gt'], transpose_ratio=0.5),
dict(
type='RandomNoise',
params=dict(
noise_type=['gaussian'],
noise_prob=[1],
gaussian_sigma=[sigma, sigma],
gaussian_gray_noise_prob=0),
keys=['img']),
dict(type='PackInputs')
]

val_pipeline = [
dict(
type='LoadImageFromFile',
key='img',
color_type='grayscale',
imdecode_backend='cv2'),
dict(
type='LoadImageFromFile',
key='gt',
color_type='grayscale',
imdecode_backend='cv2'),
dict(
type='RandomNoise',
params=dict(
noise_type=['gaussian'],
noise_prob=[1],
gaussian_sigma=[sigma, sigma],
gaussian_gray_noise_prob=0),
keys=['img']),
dict(type='PackInputs')
]

dataset settings

dataset_type = 'BasicImageDataset'
data_root = 'data'

train_dataloader = dict(
num_workers=4,
batch_size=1,
drop_last=True,
persistent_workers=False,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
ann_file='meta_info_DFWB8550sub_GT.txt',
metainfo=dict(dataset_type='dfwb', task_name='denoising'),
data_root=data_root + '/DFWB',
data_prefix=dict(img='', gt=''),
filename_tmpl=dict(img='{}', gt='{}'),
pipeline=train_pipeline))

val_dataloader = dict(
num_workers=4,
persistent_workers=False,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
metainfo=dict(dataset_type='set12', task_name='denoising'),
data_root=data_root + '/Set12',
data_prefix=dict(img='', gt=''),
pipeline=val_pipeline))

val_evaluator = [
dict(type='PSNR', prefix='Set12'),
dict(type='SSIM', prefix='Set12'),
]

train_cfg = dict(
type='IterBasedTrainLoop', max_iters=1_600_000, val_interval=5000)
val_cfg = dict(type='ValLoop')

optimizer

optim_wrapper = dict(
constructor='DefaultOptimWrapperConstructor',
type='OptimWrapper',
optimizer=dict(type='Adam', lr=2e-4, betas=(0.9, 0.999)))

learning policy

param_scheduler = dict(
type='MultiStepLR',
by_epoch=False,
milestones=[800000, 1200000, 1400000, 1500000, 1600000],
gamma=0.5)

为什么加噪声的时候sigma=15，噪声就把原来的图全部覆盖了？
从左到右分别是 原图， 输入图片，模型输出图片