ryanxingql / powervqe Goto Github PK
View Code? Open in Web Editor NEWEmpower the quality enhancement approaches for compressed videos.
License: Apache License 2.0
Empower the quality enhancement approaches for compressed videos.
License: Apache License 2.0
I copied the sources from the mmediting
folder to the local mmediting
folder from the official github repo, downloaded the stdf_ldv_v2.zip from the releases, ran the restoration_video_demo.py
, and got the error:
[/content/mmediting/mmedit/models/restorers/basic_restorer.py](https://localhost:8080/#) in forward(self, lq, gt, test_mode, **kwargs)
73
74 if test_mode:
---> 75 return self.forward_test(lq, gt, **kwargs)
76
77 return self.forward_train(lq, gt)
[/content/mmediting/mmedit/models/restorers/stdf.py](https://localhost:8080/#) in forward_test(self, lq, gt, meta, save_image, save_path, iteration)
59 dict: Output results.
60 """
---> 61 output = self.generator(lq)
62 if self.test_cfg is not None and self.test_cfg.get('metrics', None):
63 assert gt is not None, (
[/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py](https://localhost:8080/#) in _call_impl(self, *input, **kwargs)
1108 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
1109 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1110 return forward_call(*input, **kwargs)
1111 # Do not call functions when jit is used
1112 full_backward_hooks, non_full_backward_hooks = [], []
[/content/mmediting/mmedit/models/backbones/sr_backbones/stdf.py](https://localhost:8080/#) in forward(self, x)
223 Tensor: Out center frame with shape (n, c, h, w).
224 """
--> 225 out = self.stdf(x)
226 out = self.qenet(out)
227
[/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py](https://localhost:8080/#) in _call_impl(self, *input, **kwargs)
1108 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
1109 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1110 return forward_call(*input, **kwargs)
1111 # Do not call functions when jit is used
1112 full_backward_hooks, non_full_backward_hooks = [], []
[/content/mmediting/mmedit/models/backbones/sr_backbones/stdf.py](https://localhost:8080/#) in forward(self, x)
100 up_conv = getattr(self, 'up_conv{}'.format(i))
101 print(i, out.shape, out_lst[i].shape)
--> 102 out = up_conv(torch.cat([out, out_lst[i]], 1))
103
104 # compute offset and mask
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 136 but got size 135 for tensor number 1 in the list.
The out.shape
is torch.Size([1, 32, 136, 90])
while out_lst[i].shape
is torch.Size([1, 32, 135, 90])
.
I put window_size=7
in the args to run the video restoration demo. And changed the config file and the model path, of course.
Using the latest mmcv, though I'm sure the error is in sr_backbones/stdf.py
.
您好,请问图像数据集里的数据对是将HEVC应用到了图像数据上得到不同QP的压缩图像的?如何生成这些lq图像?
期待您的答复!
你好,我这有一个问题就是训练basicvsr++的时候当程序在5000次迭代验证后,出现训练速度变慢的情况,变慢了很多。在验证前的训练速度大概是2分钟左右100次迭代,但是验证结束后模型训练速度变得慢了很多。大概100次迭代要3个小时左右。请问您知道原因是什么吗。以下是我的训练日志。
20230608_115628.log
I am Vansin, the technical operator of OpenMMLab. In September of last year, we announced the release of OpenMMLab 2.0 at the World Artificial Intelligence Conference in Shanghai. We invite you to upgrade your algorithm library to OpenMMLab 2.0 using MMEngine, which can be used for both research and commercial purposes. If you have any questions, please feel free to join us on the OpenMMLab Discord at https://discord.gg/amFNsyUBvm or add me on WeChat (van-sin) and I will invite you to the OpenMMLab WeChat group.
Here are the OpenMMLab 2.0 repos branches:
OpenMMLab 1.0 branch | OpenMMLab 2.0 branch | |
---|---|---|
MMEngine | 0.x | |
MMCV | 1.x | 2.x |
MMDetection | 0.x 、1.x、2.x | 3.x |
MMAction2 | 0.x | 1.x |
MMClassification | 0.x | 1.x |
MMSegmentation | 0.x | 1.x |
MMDetection3D | 0.x | 1.x |
MMEditing | 0.x | 1.x |
MMPose | 0.x | 1.x |
MMDeploy | 0.x | 1.x |
MMTracking | 0.x | 1.x |
MMOCR | 0.x | 1.x |
MMRazor | 0.x | 1.x |
MMSelfSup | 0.x | 1.x |
MMRotate | 1.x | 1.x |
MMYOLO | 0.x |
Attention: please create a new virtual environment for OpenMMLab 2.0.
Thanks for putting the effort to standardize the framework.
I noticed you have already pushed code to mfqev2 config to be used in mmediting. Are there trained models (.pth
) I could use?
Thanks.
您好,非常感谢您的工作!我想请教下你们在竞赛里是在rgb上训练视频的吗?我看到您这边给的两个结果一个是Y-PSNR,一个是RGB-psnr(这个是mfqev2的测试集),这两个结果都是在rgb上训练然后在y或rgb通道上测的吗?为什么stdf的表现这么差,甚至都不如mfqev2(相比于之前在y通道训练 y通道测试)。mfqev2的训练集相比ldv2.0的训练集差异大吗?我如果继续做这个方向,您建议在哪个数据集上做?
报错如下:
ValueError: You may use too small dataset and our distributed sampler cannot pad your dataset correctly. We highly recommend you to use fewer GPUs to finish your work
Can we get the mmediting directory as a subtree or submodule. I'm more familiar with submodules and I partially tried to do this but the merge wasn't clean even with your fork. If you're going to allow or suggest potential updates to mmediting, I would recommend keeping changes needed for this repo in a branch of your fork. You can then simply make a submodule that tracks that branch. It is then trivial for end users to pull changes from the main mmediting repo.
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 is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
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.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.