dasongli1 / shift-net Goto Github PK
View Code? Open in Web Editor NEWA Simple Baseline for Video Restoration with Grouped Spatial-temporal Shift
A Simple Baseline for Video Restoration with Grouped Spatial-temporal Shift
How much memory (and what GPU specs) is needed for training the small model (Ours-s in the main paper)?
Thanks!
我按照CDVD-TSP的介绍下载了数据集,与Shiftnet中报道的不同,没有./datasets/GoPro/train/blur 和 ./datasets/GoPro/train/gt
而是/dataset/GOPRO/train/GOPR0372_07_00
如何得到论文中的数据集格式
Can I see the training logs please, I think I'm having a bit of a problem with the training, I can't seem to converge when I train on the DVD dataset I'm training on a 4*24G 3090.
Hi,
Is it possible to use a batch-size (per GPU) larger than 1?
The model's forward pass currently contains statements such as,
.It seems that stage0
and stage2
should be amenable to batch processing by a simple rearrangement:
x_rearranged = rearrange(x, "b f c h w -> (b f) c h w")
.
How about stage_1
? What modifications are needed for it to operate on a batchsize > 1?
您好,请问论文中Ours的训练和测试代码,以及预训练模型是哪个?我看只有Ours-s和Ours+。
谢谢!
Hello,
I am currently working with the GOPRO dataset and I am interested in running the small model with a “one_len” of 96. I would like to inquire about the GPU memory requirements for this specific configuration. Could you please provide information on how much GPU memory is needed to run the small model with a “one_len” of 96 effectively?
Thank you in advance for your assistance.
The function default_init_weights seems to be missing from basicsr/models/archs
Is this the same as the one used in EDVR? https://github.com/xinntao/EDVR/blob/b02e63a0fc5854cad7b11a87f74601612e356eff/basicsr/models/archs/arch_util.py#L21
Can you please add a LICENSE file?
Thanks :)
Could you, please, provide script to common user to treat user's video or sequental frames set?
Have you tested the time-consuming of model, how about it? Thanks
Hello,
I run the following code on an A5000 (24G) GPU:
python3 inference/test_deblur_small.py --default_data GOPRO --one_len 96
However, an out-out-memory problem happens. I can only run successfully when altering --one_len to 30, the maximum size I can set.
Which GPU do you use for using --one_len 96?
Thank you very much!
您好,我下载了您提供的Video denoising任务下,使用DAVIS训练的ours-s pretrained model。并使用提供的测试脚本和代码进行测试,得到的结果比论文提供的PSNR低0.2左右。
请问您release的测试代码和论文中测试所用的在超参数方面有没有变化?
How to modify configuration parameters so that it can handle real-time videos?
Hello author, may I ask if your model can transfer onnx? There will be an error when I use this model to transfer onnx
Traceback (most recent call last):
File "To_onnx.py", line 65, in
torch.onnx.export(model, x, 'best.onnx', input_names=input_names, opset_version=15, output_names=output_names, verbose='False')
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/init.py", line 350, in export
return utils.export(
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/utils.py", line 163, in export
_export(
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/utils.py", line 1074, in _export
graph, params_dict, torch_out = _model_to_graph(
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/utils.py", line 731, in _model_to_graph
graph = _optimize_graph(
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/utils.py", line 308, in _optimize_graph
graph = _C._jit_pass_onnx(graph, operator_export_type)
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/init.py", line 416, in _run_symbolic_function
return utils._run_symbolic_function(*args, **kwargs)
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/utils.py", line 1406, in _run_symbolic_function
return symbolic_fn(g, *inputs, **attrs)
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/symbolic_helper.py", line 219, in wrapper
args = [
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/symbolic_helper.py", line 220, in
_parse_arg(arg, arg_desc, arg_name, fn_name) # type: ignore[assignment]
File "/home/iv/miniconda3/envs/dxw/lib/python3.8/site-packages/torch/onnx/symbolic_helper.py", line 97, in _parse_arg
raise RuntimeError(
RuntimeError: Failed to export an ONNX attribute 'onnx::Cast', since it's not constant, please try to make things (e.g., kernel size) static if possible
感谢您的工作!请问一下你们的论文中FLOPs的计算结果是在多大图片上推理得到的呢?感谢您的回答。
Thanks for your great effort, I have a question can I use this code to test on single photo?
Or Are the images must be sequential frames from videos?
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.