nagadomi
56536ac133
Add support for AuxiliaryLoss
2018-10-03 18:41:32 +09:00
nagadomi
93cd40a53c
merge random erasing
2018-10-03 18:29:05 +09:00
nagadomi
4d3d123d72
Merge branch 'dev'
2017-04-15 16:30:34 +09:00
nagadomi
bb0fc3a1d3
Add -validation_filename_split option
2017-04-15 16:29:38 +09:00
nagadomi
063474b2ea
Fix clearState
2017-04-13 17:38:18 +09:00
nagadomi
b7e116de54
Add support for multi GPU training (data parallel)
...
train.lua -gpu 1,3,4
When use multi GPU mode, nccl.torch is required.
2017-04-10 20:20:17 +09:00
nagadomi
202453d6c5
Use same distributions for train/validate in user method
2017-02-26 09:03:38 +09:00
nagadomi
cdafbf00ae
Add BCE(binary cross entropy) loss for segmentation
...
Sigmoid() output is required.
2017-02-23 08:48:00 +09:00
nagadomi
dac1b89750
Fix division by zero error in validate()
2017-02-23 08:10:25 +09:00
nagadomi
8b5ccbed08
Merge branch 'dev' of into dev
2017-02-12 17:48:44 +09:00
nagadomi
763f5ddcab
Fix crop bug in rare case
2017-02-12 17:46:07 +09:00
nagadomi
29260ede24
fix batchwise psnr
2017-02-12 02:04:23 +09:00
nagadomi
b111901cbb
add -update_criterion option for back compatible
2017-02-12 01:56:03 +09:00
nagadomi
0eccbc6555
Add support for MSE loss
2017-01-09 13:00:04 +09:00
nagadomi
43a9b58fcb
Add GCN option for user method
2016-12-25 20:17:47 +09:00
nagadomi
d2cfb8f104
Add L1 criterion. Change the criterion of updating model
2016-12-05 10:32:26 +09:00
nagadomi
bdaca16c67
perfomance tuning
2016-11-02 23:42:14 +09:00
nagadomi
b066761cdc
Add data augmentation for user method
2016-10-21 18:34:48 +09:00
nagadomi
4998078d2a
use threads.safe to avoid deadlock problem
2016-10-21 00:21:57 +09:00
nagadomi
66bff68ef1
split first
2016-10-08 17:22:16 +09:00
nagadomi
06e08253e4
Fix some multhread bug
2016-09-24 07:43:15 +09:00
nagadomi
a14e6acec3
mutex
2016-09-24 05:51:47 +09:00
nagadomi
5a3d012f4e
Add random blur method for data augmentation
2016-09-24 05:32:33 +09:00
nagadomi
451ee1407f
Stop calculate the instance loss when oracle_rate=0
2016-09-11 20:59:32 +09:00
nagadomi
33e6bc888e
Fix the number of threads
2016-09-11 20:56:56 +09:00
nagadomi
c2e4bb4380
Support for multi-thread in training
...
And remove `sys`,`image` and `graphicsmagicks.conveter` from the training code because those causes the deadlock on thread package.
2016-09-11 06:31:44 +09:00
nagadomi
7ffd8a51fd
fix message
2016-08-30 17:13:52 +09:00
nagadomi
3a70d09921
create new directory when the specified directory is not found
2016-08-30 17:13:25 +09:00
nagadomi
81df729a8a
Remove -loss option
2016-07-09 15:05:11 +09:00
nagadomi
edac608f18
Add support for user specified pairwise data for universal filter
2016-07-05 02:42:40 +09:00
nagadomi
de669d24a4
Change default -validation_crops to 200.
2016-06-20 15:56:39 +09:00
nagadomi
b8ff8c6787
Remove -gamma_correction option
2016-06-10 07:37:39 +09:00
nagadomi
01b2e6d441
Remove -upsampling_filter option
2016-06-10 07:34:11 +09:00
nagadomi
e5cfd3dfce
Add -resume option
2016-06-09 02:39:52 +09:00
nagadomi
6c758ec5c0
Correct messages
2016-06-08 07:52:38 +09:00
nagadomi
307ae40883
Add noise_scale training
2016-06-08 06:39:36 +09:00
nagadomi
d0630d3a20
individual filters and box-only support
2016-06-06 14:04:13 +09:00
nagadomi
c89fd7249a
Add learning_rate_decay
2016-06-02 10:11:15 +09:00
nagadomi
634046d5f0
Fix training mode
2016-05-29 05:50:53 +09:00
nagadomi
b96bc5d453
Use correct criterion
2016-05-28 10:56:15 +09:00
nagadomi
8088460a20
Add oracle_rate option
2016-05-27 16:54:29 +09:00
nagadomi
8fec6f1b5a
Change the learning rate decay rate
2016-05-21 09:56:26 +09:00
nagadomi
7814691cbf
Add resize_blur parameter
...
latest graphicsmagick is required
2016-05-21 09:54:12 +09:00
nagadomi
a210090033
Convert model files; Add new pretrained model
...
- Add new pretrained model to ./models/upconv_7
- Move old models to ./models/vgg_7
- Use nn.LeakyReLU instead of w2nn.LeakyReLU
- Add useful attribute to .json
New JSON attribute:
The first layer has `model_config` attribute.
It contains:
model_arch: architecture name of model. see `lib/srcnn.lua`
scale_factor: if scale_factor > 1, model:forward() changes image resolution with scale_factor.
channels: input/output channels. if channels == 3, model is RGB model.
offset: pixel size that is to be removed from output.
for example:
(scale_factor=1, offset=7, input=100x100) => output=(100-7)x(100-7)
(scale_factor=2, offset=12, input=100x100) => output=(100*2-12)x(100*2-12)
And each layer has `class_name` attribute.
2016-05-15 03:04:08 +09:00
nagadomi
51ae485cd1
Add new models
...
upconv_7 is 2.3x faster than previous model
2016-05-13 09:49:53 +09:00
nagadomi
b8088ca209
Remove the limit of learning_rate_decay
2016-04-30 13:48:24 +09:00
nagadomi
7af5c9443d
Add model option and 12 layers net
2016-04-23 09:19:03 +09:00
nagadomi
68a6d4cef5
Use MSE instead of PSNR
...
PSNR depends on the minibatch size and those group.
2016-04-17 02:08:38 +09:00
nagadomi
fa9355be7c
Fix validation metric
2016-04-16 03:52:05 +09:00
nagadomi
a92b2cb386
Fix progressbar
2016-04-16 03:45:21 +09:00