b6e036147a
Add more batch norms to FlatProcessorNet_arch
2020-04-30 11:47:21 +07:00
66e91a3d9e
Revert "Enable skip-through connections from disc to gen"
...
This reverts commit b7857f35c3
2020-04-30 11:45:07 +07:00
f027e888ed
Clear out tensorboard on job restart.
2020-04-30 11:44:53 +07:00
b7857f35c3
Enable skip-through connections from disc to gen
2020-04-30 11:30:11 +07:00
bf634fc9fa
Make resnet w/ BN discriminator use leaky relus
2020-04-30 11:28:59 +07:00
3781ea725c
Add Resnet Discriminator with BN
2020-04-29 20:51:57 +07:00
a5188bb7ca
Remover fixup code from arch_util
...
Going into it's own arch.
2020-04-29 15:17:43 +07:00
5b8a77f02c
Discriminator part 1
...
New discriminator. Includes spectral norming.
2020-04-28 23:00:29 +07:00
2c145c39b6
Misc changes
2020-04-28 11:50:16 +07:00
46f550e42b
Change downsample_dataset to do no image modification
...
I'm preprocessing the images myself now. There's no need to have
the dataset do this processing as well.
2020-04-28 11:50:04 +07:00
8ab595e427
Add FlatProcessorNet
...
After doing some thinking and reading on the subject, it occurred to me that
I was treating the generator like a discriminator by focusing the network
complexity at the feature levels. It makes far more sense to process each conv
level equally for the generator, hence the FlatProcessorNet in this commit. This
network borrows some of the residual pass-through logic from RRDB which makes
the gradient path exceptionally short for pretty much all model parameters and
can be trained in O1 optimization mode without overflows again.
2020-04-28 11:49:21 +07:00
b8f67418d4
Retool HighToLowResNet
...
The receptive field of the original was *really* low. This new one has a
receptive field of 36x36px patches. It also has some gradient issues
that need to be worked out
2020-04-26 01:13:42 +07:00
02ff4a57fd
Enable HighToLowResNet to do a 1:1 transform
2020-04-25 21:36:32 +07:00
35bd1ecae4
Config changes for discriminator advantage run
...
Still going from high->low, discriminator discerns on low. Next up disc works on high.
2020-04-25 11:24:28 +07:00
d95808f4ef
Implement downsample GAN
...
This bad boy is for a workflow where you train a model on disjoint image sets to
downsample a "good" set of images like a "bad" set of images looks. You then
use that downsampler to generate a training set of paired images for supersampling.
2020-04-24 00:00:46 +07:00
ea54c7618a
Print error when image read fails
2020-04-23 23:59:32 +07:00
e98d92fc77
Allow test to operate on batches
2020-04-23 23:59:09 +07:00
8ead9ae183
Lots more config files
2020-04-23 23:58:27 +07:00
ea5f432f5a
Log total gen loss
2020-04-22 14:02:10 +07:00
79aff886b5
Modifications that allow developer to explicitly specify a different image set for PIX and feature losses
2020-04-22 10:11:14 +07:00
12d92dc443
Add GTLQ dataset
2020-04-22 00:40:38 +07:00
4d269fdac6
Support independent PIX dataroot
2020-04-22 00:40:13 +07:00
05aafef938
Support variant input sizes and scales
2020-04-22 00:39:55 +07:00
ebda70fcba
Fix AMP
2020-04-22 00:39:31 +07:00
f4b33b0531
Some random fixes/adjustments
2020-04-22 00:38:53 +07:00
2538ca9f33
Add my own configs
2020-04-22 00:37:54 +07:00
af5dfaa90d
Change GT_size to target_size
2020-04-22 00:37:41 +07:00
cc834bd5a3
Support >128px image squares
2020-04-21 16:32:59 +07:00
35a421f6ad
Add IDEA Project
2020-04-21 16:28:21 +07:00
4f6d3f0dfb
Enable AMP optimizations & write sample train images to folder.
2020-04-21 16:28:06 +07:00
9fc556be35
Remove bad folder
...
Breaks git on windows
2020-04-21 16:26:19 +07:00
a73b318f0f
Merge pull request #32 from imPRAGMA/master
...
pip requirements file, and optimisations to readme
2019-11-29 23:57:45 +07:00
1fb12871fd
Create requirements.txt
2019-11-24 07:48:52 +07:00
8e91d59107
Update README.md
2019-11-24 07:47:57 +07:00
3549100670
Merge pull request #24 from nlpjoe/master
...
Update DATASETS.md
2019-10-27 17:28:03 +07:00
ff8b645246
Update DATASETS.md
2019-10-27 16:46:45 +07:00
8c615f0763
Update README.md
2019-09-06 21:32:28 +07:00
a25ee9464d
test w/o GT
2019-09-01 22:20:10 +07:00
0098663b6b
SRGAN model supprots dist training
2019-09-01 22:14:29 +07:00
9d949b838e
Merge branch 'master' of github.com:open-mmlab/mmsr
2019-08-27 17:49:28 +07:00
866a858e59
add deform_conv_cuda_kernel.cu
2019-08-27 17:49:12 +07:00
caabd8e9c7
Update README.md
2019-08-26 17:34:32 +07:00
dfdaaef492
add DATASETS.md
2019-08-23 21:45:54 +07:00
037933ba66
mmsr
2019-08-23 21:42:47 +07:00
58b175161c
Initial commit
2019-08-23 21:04:30 +07:00
James Betker
Xintao
PRAGMA
Jzzhou
Chen Change Loy
Kai Chen