Closed MulitiByte closed 4 months ago
您好，我在使用RetinexFormer_LOL_v2_real.yml配置进行训练时，推理图像产生了随机的黑块，请问这是什么原因？ 以下是我的具体配置
name: Enhancement_RetinexFormer_LOL_v2_real model_type: ImageCleanModel scale: 1 num_gpu: 4 # set num_gpu: 0 for cpu mode manual_seed: 100
datasets: train: name: TrainSet type: Dataset_PairedImage dataroot_gt: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_train_gt dataroot_lq: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_train_lq geometric_augs: true
filename_tmpl: '{}' io_backend: type: disk # data loader use_shuffle: true num_worker_per_gpu: 8 batch_size_per_gpu: 16 ### ------- Training on single fixed-patch size 128x128--------- mini_batch_sizes: [8] iters: [300000] gt_size: 256 gt_sizes: [256] ### ------------------------------------------------------------ dataset_enlarge_ratio: 1 prefetch_mode: ~
val: name: ValSet type: Dataset_PairedImage dataroot_gt: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_val_gt dataroot_lq: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_val_lq io_backend: type: disk
network_g: type: RetinexFormer in_channels: 3 out_channels: 3 n_feat: 40 stage: 1 num_blocks: [1,2,2]
path: pretrain_network_g: ~ strict_load_g: true resume_state: ~
train: total_iter: 150000 warmup_iter: -1 # no warm up use_grad_clip: true
scheduler: type: CosineAnnealingRestartCyclicLR periods: [46000, 104000] restart_weights: [1,1] eta_mins: [0.0003,0.000001]
mixing_augs: mixup: true mixup_beta: 1.2 use_identity: true
optim_g: type: Adam lr: !!float 2e-4
betas: [0.9, 0.999]
pixel_opt: type: L1Loss loss_weight: 1 reduction: mean
val: window_size: 4 val_freq: !!float 1e3 save_img: true rgb2bgr: true use_image: false max_minibatch: 8
metrics: psnr: # metric name, can be arbitrary type: calculate_psnr crop_border: 0 test_y_channel: false
logger: print_freq: 500 save_checkpoint_freq: !!float 1e3 use_tb_logger: true wandb: project: low_light resume_id: ~
dist_params: backend: nccl port: 29500
您好，我在使用RetinexFormer_LOL_v2_real.yml配置进行训练时，推理图像产生了随机的黑块，请问这是什么原因？ 以下是我的具体配置
general settings
name: Enhancement_RetinexFormer_LOL_v2_real model_type: ImageCleanModel scale: 1 num_gpu: 4 # set num_gpu: 0 for cpu mode manual_seed: 100
dataset and data loader settings
datasets: train: name: TrainSet type: Dataset_PairedImage dataroot_gt: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_train_gt dataroot_lq: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_train_lq geometric_augs: true
val: name: ValSet type: Dataset_PairedImage dataroot_gt: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_val_gt dataroot_lq: /mnt/lustre/GPU4/home/wuzhiqi/workspace/lol_val_lq io_backend: type: disk
network structures
network_g: type: RetinexFormer in_channels: 3 out_channels: 3 n_feat: 40 stage: 1 num_blocks: [1,2,2]
path
path: pretrain_network_g: ~ strict_load_g: true resume_state: ~
training settings
train: total_iter: 150000 warmup_iter: -1 # no warm up use_grad_clip: true
Split 300k iterations into two cycles.
1st cycle: fixed 3e-4 LR for 92k iters.
2nd cycle: cosine annealing (3e-4 to 1e-6) for 208k iters.
scheduler: type: CosineAnnealingRestartCyclicLR periods: [46000, 104000]
restart_weights: [1,1] eta_mins: [0.0003,0.000001]
mixing_augs: mixup: true mixup_beta: 1.2 use_identity: true
optim_g: type: Adam lr: !!float 2e-4
weight_decay: !!float 1e-4
losses
pixel_opt: type: L1Loss loss_weight: 1 reduction: mean
validation settings
val: window_size: 4 val_freq: !!float 1e3 save_img: true rgb2bgr: true use_image: false max_minibatch: 8
metrics: psnr: # metric name, can be arbitrary type: calculate_psnr crop_border: 0 test_y_channel: false
logging settings
logger: print_freq: 500 save_checkpoint_freq: !!float 1e3 use_tb_logger: true wandb: project: low_light resume_id: ~
dist training settings
dist_params: backend: nccl port: 29500