Add RDN. #233
Add RDN. #233
Conversation
Codecov Report
@@ Coverage Diff @@
## master #233 +/- ##
==========================================
+ Coverage 79.93% 79.99% +0.05%
==========================================
Files 159 160 +1
Lines 7941 8002 +61
Branches 1177 1185 +8
==========================================
+ Hits 6348 6401 +53
- Misses 1449 1456 +7
- Partials 144 145 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
| self.G0 = mid_channels | ||
| self.G = growth_rate | ||
| self.D = num_blocks | ||
| self.C = num_layers |
There was a problem hiding this comment.
Such names (G0, G or D) are so ugly and confusing. Please rename it with regular terms.
There was a problem hiding this comment.
In general, the name of variables should not contain the capital words.
There was a problem hiding this comment.
Yeah, just use self.mid_channels = mid_channels
There was a problem hiding this comment.
G, D, C e.t. are variables in the paper, I'll change them in code.
| x (Tensor): Input tensor with shape (n, c, h, w). | ||
|
|
||
| Returns: | ||
| Tensor: Forward results, tensor with shape (n, c, h, w). |
There was a problem hiding this comment.
I don't think the shape does not change after concatenating
|
|
||
| Args: | ||
| in_channels (int): Channel number of inputs. | ||
| out_channels (int): Channel number of outputs. |
There was a problem hiding this comment.
docstring does not match __init__
| def __init__(self, in_channels, growth_rate, num_layers): | ||
| super(RDB, self).__init__() | ||
| self.layers = nn.Sequential(*[ | ||
| DenseLayer(in_channels + growth_rate * i, growth_rate) |
There was a problem hiding this comment.
is growth_rate an int or a float ratio?
There was a problem hiding this comment.
It is int, which is the channel number of dense layer output.
| Paper: Residual Dense Network for Image Super-Resolution | ||
| Adapted from: | ||
| https://github.com/yulunzhang/RDN.git | ||
| https://github.com/yjn870/RDN-pytorch |
There was a problem hiding this comment.
These GitHub projects don't have license files.
There was a problem hiding this comment.
Okay, need to contact them
| Default: 4. | ||
| num_layer (int): Layer number in the Residual Dense Block. | ||
| Default: 8. | ||
| growth_rate(int): Channels growth in each layer of RDB. |
There was a problem hiding this comment.
rate, a terminology that is often related to percentage, is not good to describe this number.
How about just growth, or channel_growth or growth_num, etc.
There was a problem hiding this comment.
picked channel_growth.
|
Mostly some comments on naming. Also need to rebase master :) |
|
Conflict with master now |
|
|
||
| Paper: Residual Dense Network for Image Super-Resolution | ||
|
|
||
| Adapted from '[email protected]:yjn870/RDN-pytorch.git' |
There was a problem hiding this comment.
use https://github.com/...
|
Linting fails |
* add RDN * Add docstring and test. * Tiny fix. * Tiny fix. * Add license. * Tiny Fix * Tiny Fix Co-authored-by: liyinshuo <[email protected]>
Residual Dense Network for Image Super-Resolution.