OpenMMLab Multimodal Advanced, Generative, and Intelligent Creation Toolbox. Unlock the magic 🪄: Generative-AI (AIGC), easy-to-use APIs, awsome model zoo, diffusion models, for text-to-image generation, image/video restoration/enhancement, etc.
Home Page: https://mmagic.readthedocs.io/en/latest/
License: Apache License 2.0
In your MSRGAN model, it seems the output is not normalized to [0, 1].
This will not affect the output quality so much if the output passes post-processes, however, it is not valid as the input for VGG networks as you said here:
# Assume input range is [0, 1]
If I missed the normalization processes, please point it out.
On the step3 "extract subimages" of preparing DIV2K datasets, I got an error 'different length of save_GT_folder and save_LR_folder.' Could anyone tell me what possible reasons are?
The length of save_LR_folder is 28800 which is the expected number and the length of save_GT_folder is 28620 which means 5 GT images are not extracted to subimages.
when i read the LQGT_dataset.py, i found that
if self.opt['color']: # change color space if necessary
img_GT = util.channel_convert(img_GT.shape[2], self.opt['color'], [img_GT])[0]
the img_GT.shape[2] is 3
the self.opt['color'] is RGB in train_SRResNet.yml
but the def channel_convert in util.py
if in_c == 3 and tar_type == 'gray': # BGR to gray
gray_list = [cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) for img in img_list]
return [np.expand_dims(img, axis=2) for img in gray_list]
elif in_c == 3 and tar_type == 'y': # BGR to y
y_list = [bgr2ycbcr(img, only_y=True) for img in img_list]
return [np.expand_dims(img, axis=2) for img in y_list]
elif in_c == 1 and tar_type == 'RGB': # gray/y to BGR
return [cv2.cvtColor(img, cv2.COLOR_GRAY2BGR) for img in img_list]
else:
return img_list
has no in_c=3 and tar_type == 'RGB',so it just return img_list。
Is the means when i use DIV2K, i don't need to deal with the conversion among BGR, gray and y?
Hello,thank you for your work first.I have a question about the size of lr pictures.I want to ask whether i can use 128128 hr pictures and 8x downsampled lr pictures(1616),i changed all scale_ratio correctly and when i ran create_lmdb.py,it shows lmdb.MapFullError: mdb_put: MDB_MAP_FULL: Environment mapsize limit reached I can get ../dataset/DIK2K/DIV2K800_sub_bicLRx8.lmdb and ../dataset/DIK2K/DIV2K800_sub.lmdb,but no meta_info.pkl under DIV2K800_sub_bicLRx8.lmdb.Can you tell me why?Thank you.
When i run python create_lmdb.py,it shows
Traceback (most recent call last):
File "create_lmdb.py", line 411, in
main()
File "create_lmdb.py", line 44, in main
general_image_folder(opt)
File "create_lmdb.py", line 120, in general_image_folder
txn.put(key_byte, data)
lmdb.MapFullError: mdb_put: MDB_MAP_FULL: Environment mapsize limit reached.
Dear Xintao:
I am now using your mmsr module to train my own model. Your code works excellent during training. But when it comes to testing, Your algorithm displayed out of memory error when processing images. I assume it is because the size of output images caused this error. I am wondering whether to solve this problem.
In class REDSDataset, it need to update the key of GT when border_mode is false and the center_frame_idx outside the range (0, 99), but name_a from key GT keeps constant all the time.
I use the DIV2K800 as my datasets. I am confused about the setting of niter、T_period、restarts with the batch_size=16, if i want to change the batch_size such as 32、64... , how to set these parameter niter、T_period、restarts、learning_rate?
These are all repos with at least 300 stars, maybe there are some credibility with these repos, and if so, is it possible to implement them and add them to the toolkit?
Here are benchmarks done by somebody else:
Here are collections/toolkit collections:
And here is a list of models:
Reasoning behind this test: Adoption of SR in drawings nagadomi/waifu2x#251
19-12-10 08:40:22.144 - INFO: Model [SRGANModel] is created.
19-12-10 08:40:22.144 - INFO: Start training from epoch: 0, iter: 0
/home/thrc_pyj/anaconda3/envs/torch1/lib/python3.7/site-packages/torch/nn/modules/loss.py:91: UserWarning: Using a target size (torch.Size([16, 3, 129, 129])) that is different to the input size (torch.Size([16, 3, 172, 172])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.
return F.l1_loss(input, target, reduction=self.reduction)
Traceback (most recent call last):
File "train.py", line 310, in
main()
File "train.py", line 158, in main
model.optimize_parameters(current_step)
File "/home/thrc_pyj/mmsr-master/codes/models/SRGAN_model.py", line 145, in optimize_parameters
l_g_pix = self.l_pix_w * self.cri_pix(self.fake_H, self.var_H)
File "/home/thrc_pyj/anaconda3/envs/torch1/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "/home/thrc_pyj/anaconda3/envs/torch1/lib/python3.7/site-packages/torch/nn/modules/loss.py", line 91, in forward
return F.l1_loss(input, target, reduction=self.reduction)
File "/home/thrc_pyj/anaconda3/envs/torch1/lib/python3.7/site-packages/torch/nn/functional.py", line 2230, in l1_loss
expanded_input, expanded_target = torch.broadcast_tensors(input, target)
File "/home/thrc_pyj/anaconda3/envs/torch1/lib/python3.7/site-packages/torch/functional.py", line 62, in broadcast_tensors
return torch._C._VariableFunctions.broadcast_tensors(tensors)
RuntimeError: The size of tensor a (172) must match the size of tensor b (129) at non-singleton dimension 3
I have no idea how the 172 come up. It should be 129(I set GT_size = 129).
Can anyone help me?
When I export an image sequence from VirtualDub, sometimes PNGs will have little errors. They still work, but CV2 cannot read them. For example, "libpng error: IDAT: invalid stored block lengths" and similar errors occur.
With hundreds of thousands of files, it is hard to predict or test. pngcheck for example cannot detect some of these errors and any other software can read them well.
BasicSR will completely stop when such an error happens.
The error happens inside the util.read_img function:
if env is None: # img
img = cv2.imread(path, cv2.IMREAD_UNCHANGED)
else:
img = _read_img_lmdb(env, path, size)
img = img.astype(np.float32) / 255.
imread throws the error, and astype then throws the Exception that crashes the script.
I've solved it by putting the call to util.read_img inside a try except clause inside __getitem__ and having it return the last image instead:
try:
img_GT = util.read_img(self.GT_env, GT_path, resolution)
except:
return self.__getitem__(index-1)
(the same for img_LQ as well)
But I am a Python noob and I think that events like this should get logged and I don't know how to log them. Can you implement something like this including logging so that one can read the log to see what image was faulty?
I try to train the EDVR use REDS dataset.use single GPU and "python train.py -opt=options/train/train_EDVR_woTSA_M.yml" i have change gpu_ids,batch_size,n_workers to 1.
i received a segment core dumped during the training begin . the program interrupt at line 123 in
"video_base_model.py ". i guess data parallel has some questions ,could you help me ?
I tried applying the video super resolution (EDVR) on other data, but I'm getting very weak results. The output barely seems to differ from the input in quality. Examples below (left is output, right is zoomed in input).
I tried both the EDVR_REDS_SR_L and the EDVR_Vimeo90K_SR_L models, with varying input sizes, getting similar results. Is this to be expected? I would guess given the REDS4 dataset was also mostly street scenes, it should at least perform similarly.
The code I'm using (adapted from test_Vid4_REDS4_with_GT.py and moved to the root folder of the repos. Although I tested it on the REDS4 dataset with no issues.
Test Vid4 (SR) and REDS4 (SR-clean, SR-blur, deblur-clean, deblur-compression) datasets
'''
import sys
sys.path.insert(0, 'codes')
import os
import os.path as osp
import glob
import logging
import numpy as np
import cv2
import torch
import utils.util as util
import data.util as data_util
import models.archs.EDVR_arch as EDVR_arch
#################
# configurations
#################
device = torch.device('cuda')
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
data_mode = 'sharp_bicubic' # Vid4 | sharp_bicubic | blur_bicubic | blur | blur_comp
# Vid4: SR
# REDS4: sharp_bicubic (SR-clean), blur_bicubic (SR-blur);
# blur (deblur-clean), blur_comp (deblur-compression).
stage = 1 # 1 or 2, use two stage strategy for REDS dataset.
flip_test = False
############################################################################
#### model
model_path = 'experiments/pretrained_models/EDVR_REDS_SR_L.pth'
N_in = 5 # use N_in images to restore one HR image
predeblur, HR_in = False, False
back_RBs = 40
model = EDVR_arch.EDVR(128, N_in, 8, 5, back_RBs, predeblur=predeblur, HR_in=HR_in)
test_dataset_folder = 'datasets/streetscenes'
#### evaluation
crop_border = 0
border_frame = N_in // 2 # border frames when evaluate
# temporal padding mode
if data_mode == 'Vid4' or data_mode == 'sharp_bicubic':
padding = 'new_info'
else:
padding = 'replicate'
save_imgs = True
save_folder = 'results/streetscenes'
util.mkdirs(save_folder)
util.setup_logger('base', save_folder, 'test', level=logging.INFO, screen=True, tofile=True)
logger = logging.getLogger('base')
#### log info
logger.info('Data: {} - {}'.format(data_mode, test_dataset_folder))
logger.info('Padding mode: {}'.format(padding))
logger.info('Model path: {}'.format(model_path))
logger.info('Save images: {}'.format(save_imgs))
logger.info('Flip test: {}'.format(flip_test))
#### set up the models
model.load_state_dict(torch.load(model_path), strict=True)
model.eval()
model = model.to(device)
img_path_l = sorted(glob.glob(osp.join(test_dataset_folder, '*')))
max_idx = len(img_path_l)
if save_imgs:
util.mkdirs(save_folder)
#### read LQ and GT images
imgs_LQ = data_util.read_img_seq(test_dataset_folder)
# process each image
for img_idx, img_path in enumerate(img_path_l):
print(img_idx, img_path)
img_name = osp.splitext(osp.basename(img_path))[0]
select_idx = data_util.index_generation(img_idx, max_idx, N_in, padding=padding)
print('select_idx:', select_idx)
imgs_in = imgs_LQ.index_select(0, torch.LongTensor(select_idx)).unsqueeze(0).to(device)
output = util.single_forward(model, imgs_in)
output = util.tensor2img(output.squeeze(0))
if save_imgs:
cv2.imwrite(osp.join(save_folder, '{}.png'.format(img_name)), output)
I am training an SRGAN model with 8x upscaling factor using an LQGT dataset. However the generator only upscales by a 4x factor.
Here is my config
use_tb_logger: True
model: srgan
distortion: sr
scale: 8
gpu_ids: [0, 1, 2, 3]
datasets:[
train:[
name: DIV2K
mode: LQGT
dataroot_LQ: path/to/my/data2.lmdb
dataroot_GT: path/to/my/data.lmdb
use_shuffle: True
n_workers: 6
batch_size: 32
GT_size: 128
use_flip: True
use_rot: True
color: RGB
phase: train
scale: 8
data_type: lmdb
]
val:[
name: DIV2K
mode: LQGT
dataroot_LQ: another/path.lmdb
dataroot_GT: path/to/data.lmdb
phase: val
scale: 8
data_type: lmdb
]
]
network_G:[
which_model_G: RRDBNet
in_nc: 3
out_nc: 3
nf: 64
nb: 16
upscale: 8
scale: 8
]
network_D:[
which_model_D: discriminator_vgg_128
in_nc: 3
nf: 64
]
path:[
pretrain_model_G: network_configs/RRDB_PSNR_x4.pth
strict_load: False
resume_state: None
experiments_root: /home/centos/init-scripts/mmsr-tissue-control/experiment
root: /home/centos/init-scripts
models: /home/centos/init-scripts/mmsr-tissue-control/experiment/models
training_state: /home/centos/init-scripts/mmsr-tissue-control/experiment/training_state
log: /home/centos/init-scripts/mmsr-tissue-control/experiment
val_images: /home/centos/init-scripts/mmsr-tissue-control/experiment/val_images
]
train:[
lr_G: 0.0001
weight_decay_G: 0
beta1_G: 0.9
beta2_G: 0.99
lr_D: 0.0001
weight_decay_D: 0
beta1_D: 0.9
beta2_D: 0.99
lr_scheme: MultiStepLR
niter: 400000
warmup_iter: -1
lr_steps: [50000, 100000, 200000, 300000]
lr_gamma: 0.5
pixel_criterion: l1
pixel_weight: 0.01
feature_criterion: l1
feature_weight: 1
gan_type: gan
gan_weight: 0.005
D_update_ratio: 1
D_init_iters: 0
manual_seed: 10
val_freq: 5000.0
]
logger:[
print_freq: 100
save_checkpoint_freq: 4000.0
]
is_train: True
dist: False
I added the following code to LQGT_dataset.py under SRGANModel.optimize_paramters: (starting at L140)
self.fake_H=self.netG(self.var_L)
print(self.var_L.shape)
print(self.fake_H.shape)
This yields the following output at train time
torch.Size([32, 3, 16, 16])
torch.Size([32, 3, 64, 64])
Is there a fix? Can someone help me shed light on this problem?
Traceback (most recent call last): File "deform_conv.py", line 10, in <module> from . import deform_conv_cuda ImportError: cannot import name 'deform_conv_cuda' from '__main__' (deform_conv.py)
Any help is appreciated !!
It's necessary to use more paper models to do something.
The model zoo is really restricted.
On the "Training and Testing" wiki page it says to Modify the configuration file, but it doesn't say how. I just want to use the pretrained models which I've already downloaded, and provide my own images to process. If I just try to run the test script without modifying anything, I get:
$ python train.py -opt options/train/train_ESRGAN.yml
...
FileNotFoundError: [Errno 2] No such file or directory: '../datasets/DIV2K/DIV2K800_sub.lmdb/meta_info.pkl'
How can we do this?
How many pictures is enough?I used 7000 pictures to train 400000 epoches.Is it enough?
Hello,
Thank you for sharing helpful data.
However, URL to download TOF results does not work.
Could you please fix it?
Thank you.
In the code about RaGan, I find that
` # pred_d_real = self.netD(self.var_ref)
# pred_d_fake = self.netD(self.fake_H.detach()) # detach to avoid BP to G
# l_d_real = self.cri_gan(pred_d_real - torch.mean(pred_d_fake), True)
# l_d_fake = self.cri_gan(pred_d_fake - torch.mean(pred_d_real), False)
# l_d_total = (l_d_real + l_d_fake) / 2
# l_d_total.backward()
pred_d_fake = self.netD(self.fake_H.detach()).detach()
pred_d_real = self.netD(self.var_ref)
l_d_real = self.cri_gan(pred_d_real - torch.mean(pred_d_fake), True) * 0.5
l_d_real.backward()
pred_d_fake = self.netD(self.fake_H.detach())
l_d_fake = self.cri_gan(pred_d_fake - torch.mean(pred_d_real.detach()), False) * 0.5
l_d_fake.backward()`
So I'd like to know waht is the difference.
In
https://github.com/open-mmlab/mmsr/blob/master/codes/models/SRGAN_model.py#L189
Why do you detach to the result of netD on fake image? Won't that prevent gradient from fake images from backproping to the discriminator?
could you tell me your server configuration?
e.g.:
GCC,CUDA,Ubuntu,
Hello, I found you add GradientPenaltyLoss loss function in your loss.py, i try to use similar gradient of image as loss function, but seems it is not used, doesn't it work in SR training?
Is it possible to use CPU (without GPU) for performing inferences on images using the pre-trained model? Such as for SRGAN?
Also, does mmsr work with the SRGAN model that is tagged with Tensorflow not pyTorch?
Thank you!
Sorry, I've got a problem. I train RCAN , RDN and EDSR according to your setting. Data preprocess is also done by your scripts. But I can't get the reported results in those papers. Empirically there is always 0.09dB difference. I wonder if you have this difference too. If so, Is that a problem about data preprocess? Thank you.
Hello, the official training configuration of EDVR is on 8 GPUs. If I train the model on only 2 1080Ti, what should I set the learning rate to? In other words, what is the relationship between lr and the number of GPUs?
I use my own dataset which has three big images in different size. Then I crop these images in 182 smaller images but they have 3 types of size. Then I use extract_subimages.py and achieve 14392 images with size of 480*480. When I try to use create_lmdb.py , I found an error said:
runfile('C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts/create_lmdb.py', wdir='C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts')
Reloaded modules: data, data.util, utils, utils.util
Reading image path list ...
data size per image is: 691200
Traceback (most recent call last):
File "", line 1, in
runfile('C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts/create_lmdb.py', wdir='C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts')
File "D:\anaconda3\lib\site-packages\spyder_kernels\customize\spydercustomize.py", line 668, in runfile
execfile(filename, namespace)
File "D:\anaconda3\lib\site-packages\spyder_kernels\customize\spydercustomize.py", line 108, in execfile
exec(compile(f.read(), filename, 'exec'), namespace)
File "C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts/create_lmdb.py", line 412, in
main()
File "C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts/create_lmdb.py", line 40, in main
general_image_folder(opt)
File "C:/Users/ThinkPad/Desktop/SR/mmsr-master/codes/data_scripts/create_lmdb.py", line 106, in general_image_folder
env = lmdb.open(lmdb_save_path, map_size=data_size * 10)
Error: C:/Users/ThinkPad/Desktop/SR/changping-GF2/train_sub_HR.lmdb: ���̿ռ䲻�㡣
I donnot know how to debug this error, can anybody help me?
As the title says, I'm getting significantly worse results than the paper when testing the EDVR_REDS_SR_L model on the RED4 dataset. I'm using torch 1.4 and CUDA 10.2, any ideas as to what's causing this? The output barely differs from the distorted input, just less sharp.
20-03-06 18:03:01.751 - INFO: Data: sharp_bicubic - datasets/REDS4/sharp_bicubic
20-03-06 18:03:01.753 - INFO: Padding mode: new_info
20-03-06 18:03:01.754 - INFO: Model path: experiments/pretrained_models/EDVR_REDS_SR_L.pth
20-03-06 18:03:01.755 - INFO: Save images: True
20-03-06 18:03:01.756 - INFO: Flip test: False
20-03-06 18:03:12.269 - INFO: 1 - 00000000 PSNR: 23.929593 dB
20-03-06 18:03:15.705 - INFO: 2 - 00000001 PSNR: 24.225784 dB
20-03-06 18:03:19.148 - INFO: 3 - 00000002 PSNR: 24.445537 dB
20-03-06 18:03:22.592 - INFO: 4 - 00000003 PSNR: 24.607198 dB
20-03-06 18:03:26.022 - INFO: 5 - 00000004 PSNR: 24.700680 dB
20-03-06 18:03:29.433 - INFO: 6 - 00000005 PSNR: 24.767155 dB
20-03-06 18:03:32.865 - INFO: 7 - 00000006 PSNR: 24.805477 dB
20-03-06 18:03:36.289 - INFO: 8 - 00000007 PSNR: 24.773191 dB
20-03-06 18:03:39.707 - INFO: 9 - 00000008 PSNR: 24.811415 dB
20-03-06 18:03:43.112 - INFO: 10 - 00000009 PSNR: 24.782979 dB
20-03-06 18:03:46.528 - INFO: 11 - 00000010 PSNR: 24.724220 dB
20-03-06 18:03:49.938 - INFO: 12 - 00000011 PSNR: 24.657968 dB
20-03-06 18:03:53.343 - INFO: 13 - 00000012 PSNR: 24.592155 dB
20-03-06 18:03:56.730 - INFO: 14 - 00000013 PSNR: 24.519794 dB
20-03-06 18:04:00.138 - INFO: 15 - 00000014 PSNR: 24.486599 dB
20-03-06 18:04:03.544 - INFO: 16 - 00000015 PSNR: 24.452770 dB
20-03-06 18:04:06.949 - INFO: 17 - 00000016 PSNR: 24.415719 dB
20-03-06 18:04:10.338 - INFO: 18 - 00000017 PSNR: 24.369736 dB
20-03-06 18:04:13.743 - INFO: 19 - 00000018 PSNR: 24.406705 dB
20-03-06 18:04:17.139 - INFO: 20 - 00000019 PSNR: 24.380248 dB
20-03-06 18:04:20.537 - INFO: 21 - 00000020 PSNR: 24.353236 dB
20-03-06 18:04:23.913 - INFO: 22 - 00000021 PSNR: 24.318037 dB
20-03-06 18:04:27.311 - INFO: 23 - 00000022 PSNR: 24.338691 dB
20-03-06 18:04:30.706 - INFO: 24 - 00000023 PSNR: 24.310392 dB
20-03-06 18:04:34.116 - INFO: 25 - 00000024 PSNR: 24.222937 dB
20-03-06 18:04:37.494 - INFO: 26 - 00000025 PSNR: 24.114522 dB
20-03-06 18:04:40.893 - INFO: 27 - 00000026 PSNR: 24.042259 dB
20-03-06 18:04:44.287 - INFO: 28 - 00000027 PSNR: 24.061988 dB
20-03-06 18:04:47.687 - INFO: 29 - 00000028 PSNR: 24.026679 dB
20-03-06 18:04:51.084 - INFO: 30 - 00000029 PSNR: 24.060308 dB
20-03-06 18:04:54.483 - INFO: 31 - 00000030 PSNR: 24.060362 dB
20-03-06 18:04:57.887 - INFO: 32 - 00000031 PSNR: 24.043007 dB
20-03-06 18:05:01.289 - INFO: 33 - 00000032 PSNR: 24.056479 dB
20-03-06 18:05:04.675 - INFO: 34 - 00000033 PSNR: 24.040209 dB
20-03-06 18:05:08.078 - INFO: 35 - 00000034 PSNR: 24.072033 dB
20-03-06 18:05:11.479 - INFO: 36 - 00000035 PSNR: 24.060528 dB
20-03-06 18:05:14.880 - INFO: 37 - 00000036 PSNR: 24.071944 dB
20-03-06 18:05:18.262 - INFO: 38 - 00000037 PSNR: 24.011707 dB
20-03-06 18:05:21.660 - INFO: 39 - 00000038 PSNR: 24.003839 dB
20-03-06 18:05:25.061 - INFO: 40 - 00000039 PSNR: 23.978071 dB
20-03-06 18:05:28.461 - INFO: 41 - 00000040 PSNR: 24.005552 dB
20-03-06 18:05:31.840 - INFO: 42 - 00000041 PSNR: 24.012480 dB
20-03-06 18:05:35.230 - INFO: 43 - 00000042 PSNR: 24.025633 dB
20-03-06 18:05:38.625 - INFO: 44 - 00000043 PSNR: 24.000322 dB
20-03-06 18:05:42.022 - INFO: 45 - 00000044 PSNR: 24.001640 dB
20-03-06 18:05:45.419 - INFO: 46 - 00000045 PSNR: 24.015714 dB
20-03-06 18:05:48.800 - INFO: 47 - 00000046 PSNR: 24.007861 dB
20-03-06 18:05:52.195 - INFO: 48 - 00000047 PSNR: 23.987225 dB
20-03-06 18:05:55.589 - INFO: 49 - 00000048 PSNR: 24.020262 dB
20-03-06 18:05:58.977 - INFO: 50 - 00000049 PSNR: 23.988086 dB
20-03-06 18:06:02.382 - INFO: 51 - 00000050 PSNR: 23.966482 dB
20-03-06 18:06:05.775 - INFO: 52 - 00000051 PSNR: 23.958009 dB
20-03-06 18:06:09.174 - INFO: 53 - 00000052 PSNR: 23.931918 dB
20-03-06 18:06:12.558 - INFO: 54 - 00000053 PSNR: 23.911961 dB
20-03-06 18:06:15.966 - INFO: 55 - 00000054 PSNR: 23.952002 dB
20-03-06 18:06:19.368 - INFO: 56 - 00000055 PSNR: 23.922195 dB
20-03-06 18:06:22.772 - INFO: 57 - 00000056 PSNR: 23.982604 dB
20-03-06 18:06:26.148 - INFO: 58 - 00000057 PSNR: 24.002386 dB
20-03-06 18:06:29.552 - INFO: 59 - 00000058 PSNR: 24.033867 dB
20-03-06 18:06:32.951 - INFO: 60 - 00000059 PSNR: 24.128416 dB
20-03-06 18:06:36.352 - INFO: 61 - 00000060 PSNR: 24.181532 dB
20-03-06 18:06:39.742 - INFO: 62 - 00000061 PSNR: 24.152309 dB
20-03-06 18:06:43.132 - INFO: 63 - 00000062 PSNR: 24.116286 dB
20-03-06 18:06:46.544 - INFO: 64 - 00000063 PSNR: 24.056074 dB
20-03-06 18:06:49.958 - INFO: 65 - 00000064 PSNR: 24.038195 dB
20-03-06 18:06:53.354 - INFO: 66 - 00000065 PSNR: 24.012471 dB
20-03-06 18:06:56.762 - INFO: 67 - 00000066 PSNR: 24.043090 dB
20-03-06 18:07:00.160 - INFO: 68 - 00000067 PSNR: 23.978444 dB
20-03-06 18:07:03.565 - INFO: 69 - 00000068 PSNR: 23.950872 dB
20-03-06 18:07:06.943 - INFO: 70 - 00000069 PSNR: 23.934508 dB
20-03-06 18:07:10.339 - INFO: 71 - 00000070 PSNR: 23.934895 dB
20-03-06 18:07:13.734 - INFO: 72 - 00000071 PSNR: 23.947487 dB
20-03-06 18:07:17.133 - INFO: 73 - 00000072 PSNR: 23.927797 dB
20-03-06 18:07:20.514 - INFO: 74 - 00000073 PSNR: 23.945414 dB
20-03-06 18:07:23.912 - INFO: 75 - 00000074 PSNR: 23.981803 dB
20-03-06 18:07:27.319 - INFO: 76 - 00000075 PSNR: 23.928317 dB
20-03-06 18:07:30.718 - INFO: 77 - 00000076 PSNR: 23.874528 dB
20-03-06 18:07:34.101 - INFO: 78 - 00000077 PSNR: 23.848243 dB
20-03-06 18:07:37.495 - INFO: 79 - 00000078 PSNR: 23.867346 dB
20-03-06 18:07:40.903 - INFO: 80 - 00000079 PSNR: 23.811867 dB
20-03-06 18:07:44.305 - INFO: 81 - 00000080 PSNR: 23.892943 dB
20-03-06 18:07:47.687 - INFO: 82 - 00000081 PSNR: 23.884709 dB
20-03-06 18:07:51.078 - INFO: 83 - 00000082 PSNR: 23.874417 dB
20-03-06 18:07:54.486 - INFO: 84 - 00000083 PSNR: 23.899484 dB
20-03-06 18:07:57.894 - INFO: 85 - 00000084 PSNR: 23.937161 dB
20-03-06 18:08:01.281 - INFO: 86 - 00000085 PSNR: 23.944973 dB
20-03-06 18:08:04.696 - INFO: 87 - 00000086 PSNR: 23.941057 dB
20-03-06 18:08:08.097 - INFO: 88 - 00000087 PSNR: 23.914812 dB
20-03-06 18:08:11.490 - INFO: 89 - 00000088 PSNR: 23.860929 dB
20-03-06 18:08:14.878 - INFO: 90 - 00000089 PSNR: 23.849738 dB
20-03-06 18:08:18.276 - INFO: 91 - 00000090 PSNR: 23.824566 dB
20-03-06 18:08:21.671 - INFO: 92 - 00000091 PSNR: 23.775884 dB
20-03-06 18:08:25.076 - INFO: 93 - 00000092 PSNR: 23.829488 dB
20-03-06 18:08:28.467 - INFO: 94 - 00000093 PSNR: 23.818127 dB
20-03-06 18:08:31.854 - INFO: 95 - 00000094 PSNR: 23.879283 dB
20-03-06 18:08:35.263 - INFO: 96 - 00000095 PSNR: 23.917133 dB
20-03-06 18:08:38.670 - INFO: 97 - 00000096 PSNR: 24.029655 dB
20-03-06 18:08:42.054 - INFO: 98 - 00000097 PSNR: 24.083472 dB
20-03-06 18:08:45.457 - INFO: 99 - 00000098 PSNR: 24.097008 dB
20-03-06 18:08:48.862 - INFO: 100 - 00000099 PSNR: 24.168127 dB
20-03-06 18:08:48.864 - INFO: Folder 000 - Average PSNR: 24.109112 dB for 100 frames; Center PSNR: 24.109278 dB for 96 frames; Border PSNR: 24.105128 dB for 4 frames.
20-03-06 18:08:55.906 - INFO: 1 - 00000000 PSNR: 25.496437 dB
20-03-06 18:08:59.279 - INFO: 2 - 00000001 PSNR: 25.442724 dB
20-03-06 18:09:02.670 - INFO: 3 - 00000002 PSNR: 25.403935 dB
20-03-06 18:09:06.070 - INFO: 4 - 00000003 PSNR: 25.439714 dB
20-03-06 18:09:09.467 - INFO: 5 - 00000004 PSNR: 25.423509 dB
20-03-06 18:09:12.852 - INFO: 6 - 00000005 PSNR: 25.250261 dB
20-03-06 18:09:16.251 - INFO: 7 - 00000006 PSNR: 25.279118 dB
20-03-06 18:09:19.646 - INFO: 8 - 00000007 PSNR: 25.253017 dB
20-03-06 18:09:23.049 - INFO: 9 - 00000008 PSNR: 25.201422 dB
20-03-06 18:09:26.439 - INFO: 10 - 00000009 PSNR: 25.210002 dB
20-03-06 18:09:29.836 - INFO: 11 - 00000010 PSNR: 25.155375 dB
20-03-06 18:09:33.224 - INFO: 12 - 00000011 PSNR: 25.176388 dB
20-03-06 18:09:36.630 - INFO: 13 - 00000012 PSNR: 25.211421 dB
20-03-06 18:09:40.021 - INFO: 14 - 00000013 PSNR: 25.256898 dB
20-03-06 18:09:43.421 - INFO: 15 - 00000014 PSNR: 25.332483 dB
20-03-06 18:09:46.830 - INFO: 16 - 00000015 PSNR: 25.368389 dB
20-03-06 18:09:50.223 - INFO: 17 - 00000016 PSNR: 25.433773 dB
20-03-06 18:09:53.608 - INFO: 18 - 00000017 PSNR: 25.366960 dB
20-03-06 18:09:57.008 - INFO: 19 - 00000018 PSNR: 25.360995 dB
20-03-06 18:10:00.416 - INFO: 20 - 00000019 PSNR: 25.334781 dB
20-03-06 18:10:03.811 - INFO: 21 - 00000020 PSNR: 25.373366 dB
20-03-06 18:10:07.184 - INFO: 22 - 00000021 PSNR: 25.362953 dB
20-03-06 18:10:10.587 - INFO: 23 - 00000022 PSNR: 25.300059 dB
20-03-06 18:10:13.990 - INFO: 24 - 00000023 PSNR: 25.283346 dB
20-03-06 18:10:17.391 - INFO: 25 - 00000024 PSNR: 25.388829 dB
20-03-06 18:10:20.778 - INFO: 26 - 00000025 PSNR: 25.499817 dB
20-03-06 18:10:24.170 - INFO: 27 - 00000026 PSNR: 25.540620 dB
20-03-06 18:10:27.569 - INFO: 28 - 00000027 PSNR: 25.490254 dB
20-03-06 18:10:30.981 - INFO: 29 - 00000028 PSNR: 25.497905 dB
20-03-06 18:10:34.365 - INFO: 30 - 00000029 PSNR: 25.609684 dB
20-03-06 18:10:37.759 - INFO: 31 - 00000030 PSNR: 25.689139 dB
20-03-06 18:10:41.162 - INFO: 32 - 00000031 PSNR: 25.562026 dB
20-03-06 18:10:44.550 - INFO: 33 - 00000032 PSNR: 25.435401 dB
20-03-06 18:10:47.953 - INFO: 34 - 00000033 PSNR: 25.382620 dB
20-03-06 18:10:51.356 - INFO: 35 - 00000034 PSNR: 25.364130 dB
20-03-06 18:10:54.760 - INFO: 36 - 00000035 PSNR: 25.359672 dB
20-03-06 18:10:58.159 - INFO: 37 - 00000036 PSNR: 25.295513 dB
20-03-06 18:11:01.540 - INFO: 38 - 00000037 PSNR: 25.383010 dB
20-03-06 18:11:04.941 - INFO: 39 - 00000038 PSNR: 25.419653 dB
20-03-06 18:11:08.331 - INFO: 40 - 00000039 PSNR: 25.346945 dB
20-03-06 18:11:11.739 - INFO: 41 - 00000040 PSNR: 25.329836 dB
20-03-06 18:11:15.130 - INFO: 42 - 00000041 PSNR: 25.256346 dB
20-03-06 18:11:18.519 - INFO: 43 - 00000042 PSNR: 25.341827 dB
20-03-06 18:11:21.918 - INFO: 44 - 00000043 PSNR: 25.323997 dB
20-03-06 18:11:25.317 - INFO: 45 - 00000044 PSNR: 25.234923 dB
20-03-06 18:11:28.705 - INFO: 46 - 00000045 PSNR: 25.273404 dB
20-03-06 18:11:32.110 - INFO: 47 - 00000046 PSNR: 25.475188 dB
20-03-06 18:11:35.516 - INFO: 48 - 00000047 PSNR: 25.507128 dB
20-03-06 18:11:38.908 - INFO: 49 - 00000048 PSNR: 25.478002 dB
20-03-06 18:11:42.300 - INFO: 50 - 00000049 PSNR: 25.473721 dB
20-03-06 18:11:45.700 - INFO: 51 - 00000050 PSNR: 25.633425 dB
20-03-06 18:11:49.108 - INFO: 52 - 00000051 PSNR: 25.583886 dB
20-03-06 18:11:52.506 - INFO: 53 - 00000052 PSNR: 25.421922 dB
20-03-06 18:11:55.882 - INFO: 54 - 00000053 PSNR: 25.511061 dB
20-03-06 18:11:59.280 - INFO: 55 - 00000054 PSNR: 25.604386 dB
20-03-06 18:12:02.682 - INFO: 56 - 00000055 PSNR: 25.781085 dB
20-03-06 18:12:06.089 - INFO: 57 - 00000056 PSNR: 25.876166 dB
20-03-06 18:12:09.490 - INFO: 58 - 00000057 PSNR: 25.761648 dB
20-03-06 18:12:12.885 - INFO: 59 - 00000058 PSNR: 25.589340 dB
20-03-06 18:12:16.289 - INFO: 60 - 00000059 PSNR: 25.636074 dB
20-03-06 18:12:19.688 - INFO: 61 - 00000060 PSNR: 25.647902 dB
20-03-06 18:12:23.077 - INFO: 62 - 00000061 PSNR: 25.523215 dB
20-03-06 18:12:26.467 - INFO: 63 - 00000062 PSNR: 25.288087 dB
20-03-06 18:12:29.876 - INFO: 64 - 00000063 PSNR: 25.228483 dB
20-03-06 18:12:33.286 - INFO: 65 - 00000064 PSNR: 25.275530 dB
20-03-06 18:12:36.668 - INFO: 66 - 00000065 PSNR: 25.303101 dB
20-03-06 18:12:40.064 - INFO: 67 - 00000066 PSNR: 25.351401 dB
20-03-06 18:12:43.456 - INFO: 68 - 00000067 PSNR: 25.221637 dB
20-03-06 18:12:46.857 - INFO: 69 - 00000068 PSNR: 25.458500 dB
20-03-06 18:12:50.236 - INFO: 70 - 00000069 PSNR: 25.631566 dB
20-03-06 18:12:53.632 - INFO: 71 - 00000070 PSNR: 25.561550 dB
20-03-06 18:12:57.034 - INFO: 72 - 00000071 PSNR: 25.613213 dB
20-03-06 18:13:00.438 - INFO: 73 - 00000072 PSNR: 25.765877 dB
20-03-06 18:13:03.825 - INFO: 74 - 00000073 PSNR: 25.791813 dB
20-03-06 18:13:07.210 - INFO: 75 - 00000074 PSNR: 25.851771 dB
20-03-06 18:13:10.607 - INFO: 76 - 00000075 PSNR: 25.857999 dB
20-03-06 18:13:14.004 - INFO: 77 - 00000076 PSNR: 25.921347 dB
20-03-06 18:13:17.398 - INFO: 78 - 00000077 PSNR: 26.144947 dB
20-03-06 18:13:20.805 - INFO: 79 - 00000078 PSNR: 26.311829 dB
20-03-06 18:13:24.217 - INFO: 80 - 00000079 PSNR: 26.209486 dB
20-03-06 18:13:27.627 - INFO: 81 - 00000080 PSNR: 26.196602 dB
20-03-06 18:13:31.016 - INFO: 82 - 00000081 PSNR: 26.160549 dB
20-03-06 18:13:34.412 - INFO: 83 - 00000082 PSNR: 26.068626 dB
20-03-06 18:13:37.800 - INFO: 84 - 00000083 PSNR: 26.142669 dB
20-03-06 18:13:41.198 - INFO: 85 - 00000084 PSNR: 26.047760 dB
20-03-06 18:13:44.576 - INFO: 86 - 00000085 PSNR: 25.821028 dB
20-03-06 18:13:47.981 - INFO: 87 - 00000086 PSNR: 25.771803 dB
20-03-06 18:13:51.383 - INFO: 88 - 00000087 PSNR: 25.808000 dB
20-03-06 18:13:54.783 - INFO: 89 - 00000088 PSNR: 25.791005 dB
20-03-06 18:13:58.171 - INFO: 90 - 00000089 PSNR: 25.701665 dB
20-03-06 18:14:01.561 - INFO: 91 - 00000090 PSNR: 25.683236 dB
20-03-06 18:14:04.967 - INFO: 92 - 00000091 PSNR: 25.732255 dB
20-03-06 18:14:08.370 - INFO: 93 - 00000092 PSNR: 25.671353 dB
20-03-06 18:14:11.762 - INFO: 94 - 00000093 PSNR: 25.553493 dB
20-03-06 18:14:15.162 - INFO: 95 - 00000094 PSNR: 25.469777 dB
20-03-06 18:14:18.564 - INFO: 96 - 00000095 PSNR: 25.585506 dB
20-03-06 18:14:21.962 - INFO: 97 - 00000096 PSNR: 25.584472 dB
20-03-06 18:14:25.350 - INFO: 98 - 00000097 PSNR: 25.433967 dB
20-03-06 18:14:28.756 - INFO: 99 - 00000098 PSNR: 25.442847 dB
20-03-06 18:14:32.141 - INFO: 100 - 00000099 PSNR: 25.417109 dB
20-03-06 18:14:32.143 - INFO: Folder 011 - Average PSNR: 25.531189 dB for 100 frames; Center PSNR: 25.534581 dB for 96 frames; Border PSNR: 25.449779 dB for 4 frames.
20-03-06 18:14:38.918 - INFO: 1 - 00000000 PSNR: 27.470123 dB
20-03-06 18:14:42.300 - INFO: 2 - 00000001 PSNR: 27.459832 dB
20-03-06 18:14:45.691 - INFO: 3 - 00000002 PSNR: 27.535595 dB
20-03-06 18:14:49.103 - INFO: 4 - 00000003 PSNR: 27.679315 dB
20-03-06 18:14:52.503 - INFO: 5 - 00000004 PSNR: 27.961649 dB
20-03-06 18:14:55.890 - INFO: 6 - 00000005 PSNR: 28.221006 dB
20-03-06 18:14:59.286 - INFO: 7 - 00000006 PSNR: 28.499709 dB
20-03-06 18:15:02.675 - INFO: 8 - 00000007 PSNR: 28.786632 dB
20-03-06 18:15:06.077 - INFO: 9 - 00000008 PSNR: 29.084916 dB
20-03-06 18:15:09.454 - INFO: 10 - 00000009 PSNR: 29.074137 dB
20-03-06 18:15:12.848 - INFO: 11 - 00000010 PSNR: 28.775812 dB
20-03-06 18:15:16.239 - INFO: 12 - 00000011 PSNR: 28.635073 dB
20-03-06 18:15:19.640 - INFO: 13 - 00000012 PSNR: 28.637493 dB
20-03-06 18:15:23.022 - INFO: 14 - 00000013 PSNR: 28.454468 dB
20-03-06 18:15:26.414 - INFO: 15 - 00000014 PSNR: 28.349233 dB
20-03-06 18:15:29.812 - INFO: 16 - 00000015 PSNR: 28.252516 dB
20-03-06 18:15:33.220 - INFO: 17 - 00000016 PSNR: 27.986286 dB
20-03-06 18:15:36.606 - INFO: 18 - 00000017 PSNR: 27.827977 dB
20-03-06 18:15:39.997 - INFO: 19 - 00000018 PSNR: 27.873088 dB
20-03-06 18:15:43.392 - INFO: 20 - 00000019 PSNR: 27.834141 dB
20-03-06 18:15:46.783 - INFO: 21 - 00000020 PSNR: 27.804338 dB
20-03-06 18:15:50.170 - INFO: 22 - 00000021 PSNR: 27.722360 dB
20-03-06 18:15:53.566 - INFO: 23 - 00000022 PSNR: 27.636687 dB
20-03-06 18:15:56.968 - INFO: 24 - 00000023 PSNR: 27.631432 dB
20-03-06 18:16:00.356 - INFO: 25 - 00000024 PSNR: 27.652337 dB
20-03-06 18:16:03.737 - INFO: 26 - 00000025 PSNR: 27.687462 dB
20-03-06 18:16:07.139 - INFO: 27 - 00000026 PSNR: 27.606636 dB
20-03-06 18:16:10.534 - INFO: 28 - 00000027 PSNR: 27.589449 dB
20-03-06 18:16:13.931 - INFO: 29 - 00000028 PSNR: 27.568649 dB
20-03-06 18:16:17.309 - INFO: 30 - 00000029 PSNR: 27.561273 dB
20-03-06 18:16:20.705 - INFO: 31 - 00000030 PSNR: 27.498807 dB
20-03-06 18:16:24.104 - INFO: 32 - 00000031 PSNR: 27.453880 dB
20-03-06 18:16:27.499 - INFO: 33 - 00000032 PSNR: 27.390142 dB
20-03-06 18:16:30.878 - INFO: 34 - 00000033 PSNR: 27.364175 dB
20-03-06 18:16:34.277 - INFO: 35 - 00000034 PSNR: 27.292243 dB
20-03-06 18:16:37.671 - INFO: 36 - 00000035 PSNR: 27.270528 dB
20-03-06 18:16:41.079 - INFO: 37 - 00000036 PSNR: 27.229522 dB
20-03-06 18:16:44.466 - INFO: 38 - 00000037 PSNR: 27.254080 dB
20-03-06 18:16:47.849 - INFO: 39 - 00000038 PSNR: 27.250369 dB
20-03-06 18:16:51.239 - INFO: 40 - 00000039 PSNR: 27.327918 dB
20-03-06 18:16:54.639 - INFO: 41 - 00000040 PSNR: 27.332572 dB
20-03-06 18:16:58.020 - INFO: 42 - 00000041 PSNR: 27.347434 dB
20-03-06 18:17:01.426 - INFO: 43 - 00000042 PSNR: 27.369476 dB
20-03-06 18:17:04.820 - INFO: 44 - 00000043 PSNR: 27.354184 dB
20-03-06 18:17:08.226 - INFO: 45 - 00000044 PSNR: 27.323591 dB
20-03-06 18:17:11.614 - INFO: 46 - 00000045 PSNR: 27.307184 dB
20-03-06 18:17:15.006 - INFO: 47 - 00000046 PSNR: 27.292381 dB
20-03-06 18:17:18.402 - INFO: 48 - 00000047 PSNR: 27.259897 dB
20-03-06 18:17:21.801 - INFO: 49 - 00000048 PSNR: 27.349362 dB
20-03-06 18:17:25.179 - INFO: 50 - 00000049 PSNR: 27.324476 dB
20-03-06 18:17:28.579 - INFO: 51 - 00000050 PSNR: 27.380226 dB
20-03-06 18:17:31.961 - INFO: 52 - 00000051 PSNR: 27.478663 dB
20-03-06 18:17:35.368 - INFO: 53 - 00000052 PSNR: 27.559490 dB
20-03-06 18:17:38.755 - INFO: 54 - 00000053 PSNR: 27.645697 dB
20-03-06 18:17:42.145 - INFO: 55 - 00000054 PSNR: 27.755224 dB
20-03-06 18:17:45.525 - INFO: 56 - 00000055 PSNR: 27.857349 dB
20-03-06 18:17:48.922 - INFO: 57 - 00000056 PSNR: 27.904337 dB
20-03-06 18:17:52.298 - INFO: 58 - 00000057 PSNR: 28.098883 dB
20-03-06 18:17:55.687 - INFO: 59 - 00000058 PSNR: 28.257135 dB
20-03-06 18:17:59.083 - INFO: 60 - 00000059 PSNR: 28.376277 dB
20-03-06 18:18:02.476 - INFO: 61 - 00000060 PSNR: 28.393124 dB
20-03-06 18:18:05.866 - INFO: 62 - 00000061 PSNR: 28.449173 dB
20-03-06 18:18:09.251 - INFO: 63 - 00000062 PSNR: 28.522042 dB
20-03-06 18:18:12.642 - INFO: 64 - 00000063 PSNR: 28.630126 dB
20-03-06 18:18:16.037 - INFO: 65 - 00000064 PSNR: 28.698283 dB
20-03-06 18:18:19.427 - INFO: 66 - 00000065 PSNR: 28.758171 dB
20-03-06 18:18:22.814 - INFO: 67 - 00000066 PSNR: 28.975227 dB
20-03-06 18:18:26.219 - INFO: 68 - 00000067 PSNR: 29.057063 dB
20-03-06 18:18:29.606 - INFO: 69 - 00000068 PSNR: 29.021664 dB
20-03-06 18:18:32.979 - INFO: 70 - 00000069 PSNR: 28.978265 dB
20-03-06 18:18:36.369 - INFO: 71 - 00000070 PSNR: 28.796016 dB
20-03-06 18:18:39.763 - INFO: 72 - 00000071 PSNR: 28.659953 dB
20-03-06 18:18:43.155 - INFO: 73 - 00000072 PSNR: 28.480837 dB
20-03-06 18:18:46.527 - INFO: 74 - 00000073 PSNR: 28.325758 dB
20-03-06 18:18:49.916 - INFO: 75 - 00000074 PSNR: 28.267757 dB
20-03-06 18:18:53.307 - INFO: 76 - 00000075 PSNR: 28.254315 dB
20-03-06 18:18:56.708 - INFO: 77 - 00000076 PSNR: 28.263080 dB
20-03-06 18:19:00.090 - INFO: 78 - 00000077 PSNR: 28.256302 dB
20-03-06 18:19:03.474 - INFO: 79 - 00000078 PSNR: 28.407808 dB
20-03-06 18:19:06.860 - INFO: 80 - 00000079 PSNR: 28.383873 dB
20-03-06 18:19:10.266 - INFO: 81 - 00000080 PSNR: 28.417908 dB
20-03-06 18:19:13.649 - INFO: 82 - 00000081 PSNR: 28.410276 dB
20-03-06 18:19:17.042 - INFO: 83 - 00000082 PSNR: 28.438092 dB
20-03-06 18:19:20.437 - INFO: 84 - 00000083 PSNR: 28.434793 dB
20-03-06 18:19:23.825 - INFO: 85 - 00000084 PSNR: 28.463180 dB
20-03-06 18:19:27.205 - INFO: 86 - 00000085 PSNR: 28.462024 dB
20-03-06 18:19:30.605 - INFO: 87 - 00000086 PSNR: 28.514308 dB
20-03-06 18:19:33.995 - INFO: 88 - 00000087 PSNR: 28.508395 dB
20-03-06 18:19:37.399 - INFO: 89 - 00000088 PSNR: 28.465283 dB
20-03-06 18:19:40.784 - INFO: 90 - 00000089 PSNR: 28.444472 dB
20-03-06 18:19:44.169 - INFO: 91 - 00000090 PSNR: 28.347403 dB
20-03-06 18:19:47.564 - INFO: 92 - 00000091 PSNR: 28.329328 dB
20-03-06 18:19:50.960 - INFO: 93 - 00000092 PSNR: 28.309115 dB
20-03-06 18:19:54.343 - INFO: 94 - 00000093 PSNR: 28.184554 dB
20-03-06 18:19:57.728 - INFO: 95 - 00000094 PSNR: 28.037867 dB
20-03-06 18:20:01.127 - INFO: 96 - 00000095 PSNR: 28.007565 dB
20-03-06 18:20:04.527 - INFO: 97 - 00000096 PSNR: 27.932315 dB
20-03-06 18:20:07.912 - INFO: 98 - 00000097 PSNR: 27.888368 dB
20-03-06 18:20:11.310 - INFO: 99 - 00000098 PSNR: 27.795131 dB
20-03-06 18:20:14.690 - INFO: 100 - 00000099 PSNR: 27.696820 dB
20-03-06 18:20:14.691 - INFO: Folder 015 - Average PSNR: 28.024312 dB for 100 frames; Center PSNR: 28.041763 dB for 96 frames; Border PSNR: 27.605476 dB for 4 frames.
20-03-06 18:20:21.707 - INFO: 1 - 00000000 PSNR: 24.895193 dB
20-03-06 18:20:25.096 - INFO: 2 - 00000001 PSNR: 24.772998 dB
20-03-06 18:20:28.495 - INFO: 3 - 00000002 PSNR: 25.506277 dB
20-03-06 18:20:31.895 - INFO: 4 - 00000003 PSNR: 25.425155 dB
20-03-06 18:20:35.285 - INFO: 5 - 00000004 PSNR: 25.719809 dB
20-03-06 18:20:38.669 - INFO: 6 - 00000005 PSNR: 25.345092 dB
20-03-06 18:20:42.064 - INFO: 7 - 00000006 PSNR: 25.466664 dB
20-03-06 18:20:45.465 - INFO: 8 - 00000007 PSNR: 25.623433 dB
20-03-06 18:20:48.864 - INFO: 9 - 00000008 PSNR: 25.605980 dB
20-03-06 18:20:52.243 - INFO: 10 - 00000009 PSNR: 25.894754 dB
20-03-06 18:20:55.642 - INFO: 11 - 00000010 PSNR: 26.180955 dB
20-03-06 18:20:59.044 - INFO: 12 - 00000011 PSNR: 25.694887 dB
20-03-06 18:21:02.437 - INFO: 13 - 00000012 PSNR: 25.953317 dB
20-03-06 18:21:05.823 - INFO: 14 - 00000013 PSNR: 25.570500 dB
20-03-06 18:21:09.227 - INFO: 15 - 00000014 PSNR: 25.591169 dB
20-03-06 18:21:12.630 - INFO: 16 - 00000015 PSNR: 25.626104 dB
20-03-06 18:21:16.032 - INFO: 17 - 00000016 PSNR: 25.564216 dB
20-03-06 18:21:19.430 - INFO: 18 - 00000017 PSNR: 24.974072 dB
20-03-06 18:21:22.823 - INFO: 19 - 00000018 PSNR: 24.795991 dB
20-03-06 18:21:26.223 - INFO: 20 - 00000019 PSNR: 24.754303 dB
20-03-06 18:21:29.620 - INFO: 21 - 00000020 PSNR: 24.567376 dB
20-03-06 18:21:33.006 - INFO: 22 - 00000021 PSNR: 24.462448 dB
20-03-06 18:21:36.410 - INFO: 23 - 00000022 PSNR: 24.400096 dB
20-03-06 18:21:39.803 - INFO: 24 - 00000023 PSNR: 24.357328 dB
20-03-06 18:21:43.194 - INFO: 25 - 00000024 PSNR: 24.500556 dB
20-03-06 18:21:46.575 - INFO: 26 - 00000025 PSNR: 24.229326 dB
20-03-06 18:21:49.970 - INFO: 27 - 00000026 PSNR: 24.289610 dB
20-03-06 18:21:53.374 - INFO: 28 - 00000027 PSNR: 24.192962 dB
20-03-06 18:21:56.771 - INFO: 29 - 00000028 PSNR: 24.143465 dB
20-03-06 18:22:00.162 - INFO: 30 - 00000029 PSNR: 24.123748 dB
20-03-06 18:22:03.560 - INFO: 31 - 00000030 PSNR: 24.254387 dB
20-03-06 18:22:06.957 - INFO: 32 - 00000031 PSNR: 24.298150 dB
20-03-06 18:22:10.371 - INFO: 33 - 00000032 PSNR: 24.354832 dB
20-03-06 18:22:13.768 - INFO: 34 - 00000033 PSNR: 24.368228 dB
20-03-06 18:22:17.153 - INFO: 35 - 00000034 PSNR: 24.442048 dB
20-03-06 18:22:20.558 - INFO: 36 - 00000035 PSNR: 24.847664 dB
20-03-06 18:22:23.947 - INFO: 37 - 00000036 PSNR: 24.762442 dB
20-03-06 18:22:27.337 - INFO: 38 - 00000037 PSNR: 24.724591 dB
20-03-06 18:22:30.738 - INFO: 39 - 00000038 PSNR: 24.886180 dB
20-03-06 18:22:34.132 - INFO: 40 - 00000039 PSNR: 24.888060 dB
20-03-06 18:22:37.533 - INFO: 41 - 00000040 PSNR: 24.618176 dB
20-03-06 18:22:40.910 - INFO: 42 - 00000041 PSNR: 24.847185 dB
20-03-06 18:22:44.320 - INFO: 43 - 00000042 PSNR: 24.897215 dB
20-03-06 18:22:47.717 - INFO: 44 - 00000043 PSNR: 24.795756 dB
20-03-06 18:22:51.102 - INFO: 45 - 00000044 PSNR: 24.718249 dB
20-03-06 18:22:54.495 - INFO: 46 - 00000045 PSNR: 24.661413 dB
20-03-06 18:22:57.893 - INFO: 47 - 00000046 PSNR: 24.617365 dB
20-03-06 18:23:01.299 - INFO: 48 - 00000047 PSNR: 24.695917 dB
20-03-06 18:23:04.711 - INFO: 49 - 00000048 PSNR: 24.837785 dB
20-03-06 18:23:08.105 - INFO: 50 - 00000049 PSNR: 24.938244 dB
20-03-06 18:23:11.502 - INFO: 51 - 00000050 PSNR: 25.087631 dB
20-03-06 18:23:14.907 - INFO: 52 - 00000051 PSNR: 24.940357 dB
20-03-06 18:23:18.316 - INFO: 53 - 00000052 PSNR: 24.993852 dB
20-03-06 18:23:21.708 - INFO: 54 - 00000053 PSNR: 25.230015 dB
20-03-06 18:23:25.106 - INFO: 55 - 00000054 PSNR: 25.191716 dB
20-03-06 18:23:28.512 - INFO: 56 - 00000055 PSNR: 25.098646 dB
20-03-06 18:23:31.925 - INFO: 57 - 00000056 PSNR: 25.164447 dB
20-03-06 18:23:35.323 - INFO: 58 - 00000057 PSNR: 25.164754 dB
20-03-06 18:23:38.731 - INFO: 59 - 00000058 PSNR: 25.045999 dB
20-03-06 18:23:42.143 - INFO: 60 - 00000059 PSNR: 25.118838 dB
20-03-06 18:23:45.536 - INFO: 61 - 00000060 PSNR: 25.179025 dB
20-03-06 18:23:48.926 - INFO: 62 - 00000061 PSNR: 25.398246 dB
20-03-06 18:23:52.330 - INFO: 63 - 00000062 PSNR: 25.076574 dB
20-03-06 18:23:55.736 - INFO: 64 - 00000063 PSNR: 25.076355 dB
20-03-06 18:23:59.146 - INFO: 65 - 00000064 PSNR: 25.121365 dB
20-03-06 18:24:02.536 - INFO: 66 - 00000065 PSNR: 24.746932 dB
20-03-06 18:24:05.934 - INFO: 67 - 00000066 PSNR: 24.578176 dB
20-03-06 18:24:09.331 - INFO: 68 - 00000067 PSNR: 24.684794 dB
20-03-06 18:24:12.748 - INFO: 69 - 00000068 PSNR: 24.862068 dB
20-03-06 18:24:16.142 - INFO: 70 - 00000069 PSNR: 25.285174 dB
20-03-06 18:24:19.550 - INFO: 71 - 00000070 PSNR: 25.261960 dB
20-03-06 18:24:22.955 - INFO: 72 - 00000071 PSNR: 25.044860 dB
20-03-06 18:24:26.378 - INFO: 73 - 00000072 PSNR: 24.940960 dB
20-03-06 18:24:29.782 - INFO: 74 - 00000073 PSNR: 24.603134 dB
20-03-06 18:24:33.188 - INFO: 75 - 00000074 PSNR: 25.015553 dB
20-03-06 18:24:36.596 - INFO: 76 - 00000075 PSNR: 24.896798 dB
20-03-06 18:24:40.007 - INFO: 77 - 00000076 PSNR: 24.415356 dB
20-03-06 18:24:43.390 - INFO: 78 - 00000077 PSNR: 24.760725 dB
20-03-06 18:24:46.799 - INFO: 79 - 00000078 PSNR: 24.642908 dB
20-03-06 18:24:50.207 - INFO: 80 - 00000079 PSNR: 24.801989 dB
20-03-06 18:24:53.608 - INFO: 81 - 00000080 PSNR: 25.023462 dB
20-03-06 18:24:57.002 - INFO: 82 - 00000081 PSNR: 25.014657 dB
20-03-06 18:25:00.410 - INFO: 83 - 00000082 PSNR: 24.944733 dB
20-03-06 18:25:03.821 - INFO: 84 - 00000083 PSNR: 24.675026 dB
20-03-06 18:25:07.219 - INFO: 85 - 00000084 PSNR: 24.549828 dB
20-03-06 18:25:10.612 - INFO: 86 - 00000085 PSNR: 24.490831 dB
20-03-06 18:25:14.013 - INFO: 87 - 00000086 PSNR: 24.580773 dB
20-03-06 18:25:17.413 - INFO: 88 - 00000087 PSNR: 24.532459 dB
20-03-06 18:25:20.837 - INFO: 89 - 00000088 PSNR: 24.530201 dB
20-03-06 18:25:24.236 - INFO: 90 - 00000089 PSNR: 24.219955 dB
20-03-06 18:25:27.646 - INFO: 91 - 00000090 PSNR: 24.444490 dB
20-03-06 18:25:31.054 - INFO: 92 - 00000091 PSNR: 24.137894 dB
20-03-06 18:25:34.452 - INFO: 93 - 00000092 PSNR: 24.212179 dB
20-03-06 18:25:37.855 - INFO: 94 - 00000093 PSNR: 24.144597 dB
20-03-06 18:25:41.257 - INFO: 95 - 00000094 PSNR: 24.173200 dB
20-03-06 18:25:44.658 - INFO: 96 - 00000095 PSNR: 24.240060 dB
20-03-06 18:25:48.066 - INFO: 97 - 00000096 PSNR: 24.308548 dB
20-03-06 18:25:51.459 - INFO: 98 - 00000097 PSNR: 24.291141 dB
20-03-06 18:25:54.871 - INFO: 99 - 00000098 PSNR: 24.288709 dB
20-03-06 18:25:58.271 - INFO: 100 - 00000099 PSNR: 24.315503 dB
20-03-06 18:25:58.273 - INFO: Folder 020 - Average PSNR: 24.842291 dB for 100 frames; Center PSNR: 24.853716 dB for 96 frames; Border PSNR: 24.568101 dB for 4 frames.
20-03-06 18:25:58.274 - INFO: ################ Tidy Outputs ################
20-03-06 18:25:58.275 - INFO: Folder 000 - Average PSNR: 24.109112 dB. Center PSNR: 24.109278 dB. Border PSNR: 24.105128 dB.
20-03-06 18:25:58.276 - INFO: Folder 011 - Average PSNR: 25.531189 dB. Center PSNR: 25.534581 dB. Border PSNR: 25.449779 dB.
20-03-06 18:25:58.277 - INFO: Folder 015 - Average PSNR: 28.024312 dB. Center PSNR: 28.041763 dB. Border PSNR: 27.605476 dB.
20-03-06 18:25:58.278 - INFO: Folder 020 - Average PSNR: 24.842291 dB. Center PSNR: 24.853716 dB. Border PSNR: 24.568101 dB.
20-03-06 18:25:58.278 - INFO: ################ Final Results ################
20-03-06 18:25:58.279 - INFO: Data: sharp_bicubic - datasets/REDS4/sharp_bicubic
20-03-06 18:25:58.281 - INFO: Padding mode: new_info
20-03-06 18:25:58.282 - INFO: Model path: experiments/pretrained_models/EDVR_REDS_SR_L.pth
20-03-06 18:25:58.283 - INFO: Save images: True
20-03-06 18:25:58.283 - INFO: Flip test: False
20-03-06 18:25:58.285 - INFO: Total Average PSNR: 25.626726 dB for 4 clips. Center PSNR: 25.634834 dB. Border PSNR: 25.432121 dB.
Honestly, I believe this project will be much more helpful if it's able to train on custom datasets with only a few changes.
After running 'python train.py -opt options/train/train_ESRGAN.yml' with my own images, lmdb and yml configured accordingly, I keep getting this error:
19-10-08 08:00:18.801 - INFO: Loading model for G [../experiments/pretrained_models/RRDB_PSNR_x4.pth] ...
19-10-08 08:00:18.956 - INFO: Model [SRGANModel] is created.
19-10-08 08:00:18.956 - INFO: Start training from epoch: 0, iter: 0
Traceback (most recent call last):
File "train.py", line 310, in
main()
File "train.py", line 149, in main
for _, train_data in enumerate(train_loader):
File "/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 582, in next
return self._process_next_batch(batch)
File "/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 608, in _process_next_batch
raise batch.exc_type(batch.exc_msg)
RuntimeError: Traceback (most recent call last):
File "/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 99, in _worker_loop
samples = collate_fn([dataset[i] for i in batch_indices])
File "/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 63, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in batch[0]}
File "/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 63, in
return {key: default_collate([d[key] for d in batch]) for key in batch[0]}
File "/home/ubuntu/anaconda3/envs/pytorch_p36/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 43, in default_collate
return torch.stack(batch, 0, out=out)
RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0. Got 128 and 0 in dimension 3 at /opt/conda/conda-bld/pytorch_1556653099582/work/aten/src/TH/generic/THTensor.cpp:711
Any help is very much appreciated.
Thanks for the great project. Looking through code, I noticed that in SRGAN_model.py there are such lines:
if opt['dist']:
pass # do not need to use DistributedDataParallel for netF
else:
self.netF = DataParallel(self.netF)
can you please provide motivation of not wrapping netF with DistributedDataParallel
I love to experimenting with neural networks and machine learning.
Thank you for this great tool.
So, i read that i could put pictures in a general image folder.
Unfortunately the images folder location is not clear for me.
I got my to-enhace-Images from a movie in mmsr/images/image-xxxx.jpg
I downloaded the movie with youtube-dl and converted to still frames with ffmpeg.
python train.py -opt options/train/train_SRResNet.yml
export CUDA_VISIBLE_DEVICES=0
Disabled distributed training.
19-09-05 11:13:52.512 - INFO: name: 001_MSRResNetx4_scratch_DIV2K
use_tb_logger: True
model: sr
distortion: sr
scale: 4
gpu_ids: [0]
datasets:[
train:[
name: DIV2K
mode: LQGT
dataroot_GT: ../datasets/DIV2K/DIV2K800_sub.lmdb
dataroot_LQ: ../datasets/DIV2K/DIV2K800_sub_bicLRx4.lmdb
use_shuffle: True
n_workers: 6
batch_size: 16
GT_size: 128
use_flip: True
use_rot: True
color: RGB
phase: train
scale: 4
data_type: lmdb
]
val:[
name: val_set5
mode: LQGT
dataroot_GT: ../datasets/val_set5/Set5
dataroot_LQ: ../datasets/val_set5/Set5_bicLRx4
phase: val
scale: 4
data_type: img
]
]
network_G:[
which_model_G: MSRResNet
in_nc: 3
out_nc: 3
nf: 64
nb: 16
upscale: 4
scale: 4
]
path:[
pretrain_model_G: None
strict_load: True
resume_state: None
root: /content/mmsr
experiments_root: /content/mmsr/experiments/001_MSRResNetx4_scratch_DIV2K
models: /content/mmsr/experiments/001_MSRResNetx4_scratch_DIV2K/models
training_state: /content/mmsr/experiments/001_MSRResNetx4_scratch_DIV2K/training_state
log: /content/mmsr/experiments/001_MSRResNetx4_scratch_DIV2K
val_images: /content/mmsr/experiments/001_MSRResNetx4_scratch_DIV2K/val_images
]
train:[
lr_G: 0.0002
lr_scheme: CosineAnnealingLR_Restart
beta1: 0.9
beta2: 0.99
niter: 1000000
warmup_iter: -1
T_period: [250000, 250000, 250000, 250000]
restarts: [250000, 500000, 750000]
restart_weights: [1, 1, 1]
eta_min: 1e-07
pixel_criterion: l1
pixel_weight: 1.0
manual_seed: 10
val_freq: 5000.0
]
logger:[
print_freq: 100
save_checkpoint_freq: 5000.0
]
is_train: True
dist: False19-09-05 11:13:54.299 - INFO: Random seed: 10
WARNING: Logging before flag parsing goes to stderr.
I0905 11:13:54.299421 140545874597760 train.py:94] Random seed: 10
Traceback (most recent call last):
File "train.py", line 310, in
main()
File "train.py", line 104, in main
train_set = create_dataset(dataset_opt)
File "/content/mmsr/codes/data/init.py", line 44, in create_dataset
dataset = D(dataset_opt)
File "/content/mmsr/codes/data/LQGT_dataset.py", line 24, in init
self.paths_GT, self.sizes_GT = util.get_image_paths(self.data_type, opt['dataroot_GT'])
File "/content/mmsr/codes/data/util.py", line 51, in get_image_paths
paths, sizes = _get_paths_from_lmdb(dataroot)
File "/content/mmsr/codes/data/util.py", line 37, in _get_paths_from_lmdb
meta_info = pickle.load(open(os.path.join(dataroot, 'meta_info.pkl'), 'rb'))
FileNotFoundError: [Errno 2] No such file or directory: '../datasets/DIV2K/DIV2K800_sub.lmdb/meta_info.pkl'
So i create a lmdb, which would not actually be necessary form your description on Datasets,
the Script data_scipts/create_lmdb.py looks in DIV2K800_sub and fails too.
python data_scipts/create_lmdb.py
Reading image path list ...
Traceback (most recent call last):
File "data_scripts/create_lmdb.py", line 411, in
main()
File "data_scripts/create_lmdb.py", line 39, in main
general_image_folder(opt)
File "data_scripts/create_lmdb.py", line 102, in general_image_folder
data_size_per_img = cv2.imread(all_img_list[0], cv2.IMREAD_UNCHANGED).nbytes
IndexError: list index out of range
Hi,
Thank you for providing a powerful tool. But sometime the input data maybe is not pictures, Is it possible to realise a new dataloader without lmdb in your framework and which part of your framework need to be changed expect dataload?
The test image and the corresponding GT must be defined in test_ESRGAN.yml. If I just want to test a random image, this image has no GT, what should I do?
src/deform_conv_cuda_kernel.cu(258): error: no suitable constructor exists to convert from "c10::ScalarType" to "at::Type"
src/deform_conv_cuda_kernel.cu(352): error: no suitable constructor exists to convert from "c10::ScalarType" to "at::Type"
src/deform_conv_cuda_kernel.cu(450): error: no suitable constructor exists to convert from "c10::ScalarType" to "at::Type"
src/deform_conv_cuda_kernel.cu(780): error: no suitable constructor exists to convert from "c10::ScalarType" to "at::Type"
src/deform_conv_cuda_kernel.cu(812): error: no suitable constructor exists to convert from "c10::ScalarType" to "at::Type"
src/deform_conv_cuda_kernel.cu(845): error: no suitable constructor exists to convert from "c10::ScalarType" to "at::Type"
6 errors detected in the compilation of "/tmp/tmpxft_000126d0_00000000-6_deform_conv_cuda_kernel.cpp1.ii".
error: command '/home/work/cuda-9.0/bin/nvcc' failed with exit status 1
Hi,
Thanks for this wonderful repo, it's really useful in my current research.
I want to fine-tune SRGAN on a custom dataset using pre-trained weights. However, only the weights for the generators are available in the model-zoo. I was wondering if you could provide the discriminator weights as well?
Thanks in advance
Thanks for your great work!
I have some question about the following line in your code
https://github.com/open-mmlab/mmsr/blob/8c615f07638e55003dcacd0808b711d62241b20b/codes/models/SRGAN_model.py#L182
In the code，the l_d_real multiply 0.5 after calculate the loss，and pred_d_fake is defined twice, what is the different bewteen your final code and the code you have commented out
https://github.com/open-mmlab/mmsr/blob/8c615f07638e55003dcacd0808b711d62241b20b/codes/models/SRGAN_model.py#L183
pred_d_real = self.netD(self.var_ref)
pred_d_fake = self.netD(self.fake_H.detach()) # detach to avoid BP to G
l_d_real = self.cri_gan(pred_d_real - torch.mean(pred_d_fake), True)
l_d_fake = self.cri_gan(pred_d_fake - torch.mean(pred_d_real), False)
l_d_total = (l_d_real + l_d_fake) / 2
l_d_total.backward()
There are some missing files, such as deform_conv_cuda_kernel.cu. More details can be found here, https://github.com/open-mmlab/mmdetection/tree/master/mmdet/ops/dcn/src, copy this into src directory can be fixed.
Hello,
I have trained X4 using the default yml file, but I am not able to change the configuration to train X2 , X3 or X8. Could please tell me where the changes needs to be made ?
20-03-04 20:15:37.214 - INFO: [epoch: 4, iter: 4,900, lr:(1.000e-04,)] l_g_pix: 3.7640e-04 l_g_fea: 9.9916e-01 l_g_gan: 2.9927e-02 l_d_real: 1.9228e-03 l_d_fake: 1.5228e-03 D_real: 7.3041e+01 D_fake: 6.7059e+01
20-03-04 20:17:49.691 - INFO: [epoch: 4, iter: 5,000, lr:(1.000e-04,)] l_g_pix: 3.4310e-04 l_g_fea: 9.8867e-01 l_g_gan: 1.9802e-02 l_d_real: 2.2165e-02 l_d_fake: 1.5732e-02 D_real: 1.2171e+02 D_fake: 1.1779e+02
[ ] 0/4144, elapsed: 0s, ETA:
Start...
Traceback (most recent call last):
File "train.py", line 310, in
mse = np.mean((img1 - img2)**2)
main()
File "train.py", line 202, in main
avg_psnr += util.calculate_psnr(sr_img, gt_img)
File "/workdir/mmsr/codes/utils/util.py", line 230, in calculate_psnr
ValueError: operands could not be broadcast together with shapes (1912,1912,3) (112,112,3)
hello,I train the model use MSRResNet without using gan,but the loss only fluctuates but doesn't down ,
while the psnr increases
The lr_scheme that I use is CosineAnnealingLR_Restart
the tensorboard out :
when
when
Hello
I'm trying to train ESRGAN on 2 GPU (RTX 2080 TI). Training crash after this error:
20-03-14 20:58:49.211 - INFO: [epoch: 0, iter: 4,900, lr:(1.000e-04,)] l_g_pix: 5.1419e-04 l_g_fea: 1.7966e+00 l_g_gan: 6.4983e-02 l_d_real: 1.4380e-06 l_d_fake: 1.5758e-06 D_real: 5.7167e+01 D_fake: 4.4170e+01
20-03-14 20:59:32.886 - INFO: [epoch: 0, iter: 5,000, lr:(1.000e-04,)] l_g_pix: 6.5671e-04 l_g_fea: 1.5129e+00 l_g_gan: 2.0994e-02 l_d_real: 1.3033e-02 l_d_fake: 1.0278e-02 D_real: 3.3355e+01 D_fake: 2.9180e+01
[ ] 0/100, elapsed: 0s, ETA:
Start...
Traceback (most recent call last):
File "train.py", line 310, in
main()
File "train.py", line 220, in main
val_data['LQs'].unsqueeze_(0)
KeyError: 'LQs'
�[2F�[J[------------------------------] 1/100, 0.3 task/s, elapsed: 3s, ETA: 305s
Test 0801
�[2F�[J[------------------------------] 2/100, 0.5 task/s, elapsed: 4s, ETA: 194s
Test 0802
�[2F�[J[------------------------------] 3/100, 0.4 task/s, elapsed: 7s, ETA: 219s
Any idea on what's causing this ?
Hi,could you please provide pretrained model D?
How to use EDVR test? Can you provide yam file for EDVR test?
19-10-11 18:15:50.311 - INFO: Data: Vid4 - ../content/drive/My Drive/app/mmsr/datasets/Vid4/BIx4/city/*
19-10-11 18:15:50.312 - INFO: Padding mode: new_info
19-10-11 18:15:50.312 - INFO: Model path: /content/drive/My Drive/app/mmsr/experiments/pretrained_models/EDVR_Vimeo90K_SR_L.pth
19-10-11 18:15:50.313 - INFO: Save images: True
19-10-11 18:15:50.313 - INFO: Flip test: False
19-10-11 18:15:52.915 - INFO: ################ Tidy Outputs ################
19-10-11 18:15:52.916 - INFO: ################ Final Results ################
19-10-11 18:15:52.916 - INFO: Data: Vid4 - ../content/drive/My Drive/app/mmsr/datasets/Vid4/BIx4/city/*
19-10-11 18:15:52.916 - INFO: Padding mode: new_info
19-10-11 18:15:52.916 - INFO: Model path: /content/drive/My Drive/app/mmsr/experiments/pretrained_models/EDVR_Vimeo90K_SR_L.pth
19-10-11 18:15:52.917 - INFO: Save images: True
19-10-11 18:15:52.917 - INFO: Flip test: False
Traceback (most recent call last):
File "codes/test_Vid4_REDS4_with_GT.py", line 208, in
main()
File "codes/test_Vid4_REDS4_with_GT.py", line 202, in main
sum(avg_psnr_l) / len(avg_psnr_l), len(subfolder_l),
ZeroDivisionError: division by zero
I am training EDVR with REDS dataset with your pre-trained model. But I have too big loss (8e+6 ~ 1.4e+7). Could you tell me your loss scale?
In the function calculate_ssim, it calculates ssim(img1, img2) for three times. Did you try to calculate ssim for different channels? If so, it's a bug.
def calculate_ssim(img1, img2):
'''calculate SSIM
the same outputs as MATLAB's
img1, img2: [0, 255]
'''
if not img1.shape == img2.shape:
raise ValueError('Input images must have the same dimensions.')
if img1.ndim == 2:
return ssim(img1, img2)
elif img1.ndim == 3:
if img1.shape[2] == 3:
ssims = []
for i in range(3):
ssims.append(ssim(img1, img2))
return np.array(ssims).mean()
elif img1.shape[2] == 1:
return ssim(np.squeeze(img1), np.squeeze(img2))
else:
raise ValueError('Wrong input image dimensions.')
Can i use RRDBNet instead of MSRResNet when i train srgan,will they share the same result?I saw that they will be almost same in ESRGAN.
After running 'python train.py -opt options/train/train_SRResNet.yml' with DIV2K dataset,I get this error:
`export CUDA_VISIBLE_DEVICES=0,1
Traceback (most recent call last):
File "/data/thrc_shx/PycharmProjects/mmsr-master/codes/train.py", line 312, in
main()
File "/data/thrc_shx/PycharmProjects/mmsr-master/codes/train.py", line 46, in main
init_dist()
File "/data/thrc_shx/PycharmProjects/mmsr-master/codes/train.py", line 26, in init_dist
dist.init_process_group(backend=backend, **kwargs)
File "/home/thrc_shx/anaconda3/envs/py36pytorch/lib/python3.6/site-packages/torch/distributed/distributed_c10d.py", line 406, in init_process_group
store, rank, world_size = next(rendezvous(url))
File "/home/thrc_shx/anaconda3/envs/py36pytorch/lib/python3.6/site-packages/torch/distributed/rendezvous.py", line 119, in _env_rendezvous_handler
raise _env_error("RANK")
ValueError: Error initializing torch.distributed using env:// rendezvous: environment variable RANK expected, but not set
Process finished with exit code 1`
Any help is very much appreciated.
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.