knazeri/edge-connect issues and pull requests

#196 - Edge Model Not converging

Issue - State: open - Opened by danielanojan 2 months ago

#195 - a question

Issue - State: open - Opened by shs1235 7 months ago

#194 - Replace outdated python modules with latest versions and resolved some minor issues

Pull Request - State: open - Opened by codevisioner 9 months ago - 1 comment

#193 - Export models as `ONNX` files

Pull Request - State: open - Opened by Luni-4 about 1 year ago

#192 - Fix test function

Pull Request - State: open - Opened by Luni-4 about 1 year ago

#191 - RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1.

Issue - State: open - Opened by kaelyavel over 1 year ago

#190 - The loss function is abnormal when the edge network is trained

Issue - State: open - Opened by TT-mouse over 1 year ago

#189 - About precision and recall during training

Issue - State: open - Opened by Ghost0405 over 1 year ago - 1 comment

#188 - When I tried to start training, I got an error：RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]] is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

Issue - State: open - Opened by TT-mouse over 1 year ago - 15 comments

#187 - Why is there an error when I train MODEL4: joint model/为什么我训练MODEL4 ：joint model会报错

Issue - State: open - Opened by Hgy12345 over 1 year ago

#186 - When using edge=2, training has ValueError: operands could not be broadcast together with shapes (256,256,3) (256,256)

Issue - State: open - Opened by wzy46775852 over 1 year ago

#185 - New easy to use inpanting method with transformers

Issue - State: open - Opened by mohammadrezanaderi4 over 1 year ago - 1 comment

#184 - Sizes of tensors must match except in dimension 1

Issue - State: open - Opened by tendrillion almost 2 years ago

#183 - how to implement the visualization for the learned edges?

Issue - State: open - Opened by micklexqg almost 2 years ago - 2 comments

#182 - Canny sigma

Issue - State: open - Opened by 1997Jessie about 2 years ago - 1 comment

#181 - Selection of dataset

Issue - State: open - Opened by chandniagarwal about 2 years ago

#180 - Training on Google Colab immediately stops

Issue - State: open - Opened by szymek1 about 2 years ago - 1 comment

#179 - 如果对图像修复，edge-connect感兴趣，或者需要帮助，可以联系我

Issue - State: open - Opened by anshen666 over 2 years ago

#178 - Hello, After reading your paper, may I have a question that why you choice 178 for the celebA dataset drop size.

Issue - State: open - Opened by FavorMylikes over 2 years ago

#177 - Convergency of edge model

Issue - State: open - Opened by lizijue over 2 years ago - 10 comments

#176 - Run the program on CoLab

Issue - State: open - Opened by mahdikhishvand over 2 years ago

#175 - Who can help me slove this error? (when I try to train )

Issue - State: open - Opened by lgl1227 over 2 years ago - 5 comments

#174 - Test image is being filled in a lighter shade

Issue - State: closed - Opened by rgao3871 almost 3 years ago - 1 comment

#173 - Results of first stage: edge model

Issue - State: open - Opened by Emy-cv almost 3 years ago - 7 comments

#172 - About supplementary material of your paper

Issue - State: open - Opened by ChisonShang about 3 years ago

#171 - Recall and Precision

Issue - State: open - Opened by lixiaocong-svg over 3 years ago - 6 comments

#170 - How to generate masks and edges for custom image datasets

Issue - State: open - Opened by tiwarikaran over 3 years ago

#169 - About images automatic impainting

Issue - State: open - Opened by yjdqk over 3 years ago

#168 - Executing the edge-connect model in Windows 10

Issue - State: open - Opened by Sicily-F over 3 years ago - 2 comments

#167 - Question about Paris StreetView validation set

Issue - State: open - Opened by kiet13 over 3 years ago

#166 - Unable to get any result on celeba trained model

Issue - State: open - Opened by chandniagarwal almost 4 years ago - 1 comment

#165 - About dataset process

Issue - State: open - Opened by BlossomingL almost 4 years ago

#164 - png size

Issue - State: open - Opened by yy2016417 almost 4 years ago - 1 comment

#163 - Training time

Issue - State: open - Opened by CyrilCsy almost 4 years ago

#162 - Paper about Examples of object removal and image editing using our EdgeConnect model

Issue - State: open - Opened by lantingyue11 almost 4 years ago

#161 - train error

Issue - State: open - Opened by lantingyue11 almost 4 years ago - 5 comments

#160 - How many epochs on each stage?

Issue - State: open - Opened by xiaoj45 almost 4 years ago

#159 - Paris Street View Testing Results

Issue - State: closed - Opened by yingchen001 almost 4 years ago - 1 comment

#158 - Value problem

Issue - State: open - Opened by weizitai about 4 years ago

#157 - High mae value using the pretrained model on Places2

Issue - State: open - Opened by ArlenCHEN about 4 years ago

#156 - how to get mask when repair real picture

Issue - State: open - Opened by 15838169447 about 4 years ago

#155 - Some problems about the model

Issue - State: open - Opened by juechen-zzz about 4 years ago

#154 - Weights init error

Issue - State: open - Opened by miaodd98 about 4 years ago

#153 - Pretrained weigths

Issue - State: closed - Opened by pvitoria about 4 years ago

#152 - RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]] is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

Issue - State: closed - Opened by CyrilShch over 4 years ago - 4 comments

#151 - Train images with paired masks

Issue - State: closed - Opened by Meulen92 over 4 years ago - 1 comment

#150 - YAMLLoadWarning

Issue - State: open - Opened by ArtScanner over 4 years ago - 1 comment

#149 - How to create a training directory？

Issue - State: open - Opened by ArtScanner over 4 years ago - 1 comment

#148 - What does the number in the log file mean?

Issue - State: closed - Opened by GODGANG4885 over 4 years ago

#147 - I'm trying to train new data. I wonder what the training data is like.

Issue - State: open - Opened by cjfgml123 over 4 years ago

#146 - fid score error

Issue - State: closed - Opened by miaodd98 over 4 years ago - 1 comment

#145 - Masked image from results doesn't match input

Issue - State: closed - Opened by violetamenendez over 4 years ago - 1 comment

#144 - what is the difference of model training of stage3 and stage4

Issue - State: open - Opened by 12ycli over 4 years ago - 6 comments

#143 - A question of testing

Issue - State: closed - Opened by Arbiter0 over 4 years ago - 1 comment

#142 - Why not use L1 loss between the masked parts in output img and input img.

Issue - State: open - Opened by ChenYutongTHU over 4 years ago

#141 - a mistake happened when loading the pre-trained model,why??

Issue - State: open - Opened by 582751489 over 4 years ago

#140 - keyerror

Issue - State: open - Opened by Tim-Tianyu over 4 years ago

#139 - About the implementation of perception loss

Issue - State: open - Opened by vinthony over 4 years ago

#138 - Why not use the L1_Loss during training EdgeGenerator？

Issue - State: open - Opened by laolongboy over 4 years ago

#137 - something wrong on testing

Issue - State: open - Opened by 1559145087 over 4 years ago - 5 comments

#136 - inpainting results

Issue - State: closed - Opened by tks-reply over 4 years ago

#135 - outputs = self(images, edges, masks)

Issue - State: open - Opened by WxpHB over 4 years ago

#134 - Mistakes in the forward functions in models.py

Issue - State: open - Opened by hedjm almost 5 years ago - 1 comment

#133 - 我没跑这个代码，只是使用了它训练好的CelebA模型产生些修复数据

Issue - State: open - Opened by zhengbowei almost 5 years ago - 3 comments

#130 - pre-trained models run errors

Issue - State: open - Opened by xiaoming3526 almost 5 years ago - 8 comments

#127 - Why should provide a mask for damaged image in test mode?

Issue - State: open - Opened by DequanZhu almost 5 years ago - 3 comments

#124 - RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation

Issue - State: open - Opened by renleidewenming almost 5 years ago - 4 comments

#121 - Missing keys in state_dict?

Issue - State: open - Opened by laolongboy almost 5 years ago - 14 comments

#120 - problems about model saving

Issue - State: open - Opened by HuaZheLei almost 5 years ago - 1 comment

#119 - A question about discriminator design

Issue - State: open - Opened by XiaoqiangZhou almost 5 years ago - 1 comment

#117 - why

Issue - State: open - Opened by donghang941114 almost 5 years ago - 1 comment

#116 - problem on testing your CelebA model

Issue - State: open - Opened by zhengbowei about 5 years ago - 1 comment

#111 - A small problem about the code in 'edge_connect.py'

Issue - State: closed - Opened by XiaoqiangZhou about 5 years ago - 3 comments

#110 - What is your recall and precision of your edge prediction network

Issue - State: closed - Opened by neouyghur about 5 years ago - 1 comment

#109 - Can not download the pretrained model

Issue - State: open - Opened by Zealoe about 5 years ago - 2 comments

#108 - edge model does not converge

Issue - State: open - Opened by Turlan about 5 years ago - 8 comments

#107 - I doubt if the input data I trained is not a problem？

Issue - State: open - Opened by 695874419 about 5 years ago - 1 comment

#106 - What are the differences between previous edge-based methods and edge-connect?

Issue - State: closed - Opened by raywzy about 5 years ago

#105 - some question about MAX_ITERS parameter.

Issue - State: open - Opened by 695874419 about 5 years ago - 1 comment

#104 - 请问我可以尝试用这个来复原下面这种图像嘛

Issue - State: open - Opened by yuxiaomu about 5 years ago - 1 comment

#103 - Runtime error while training model =1

Issue - State: open - Opened by krazzzyKITTY over 5 years ago - 2 comments

#102 - Low clarity

Issue - State: open - Opened by primejava over 5 years ago

#101 - turns on debugging mode

Issue - State: open - Opened by ashergaga over 5 years ago

#100 - How long do you trained?

Issue - State: open - Opened by 12ycli over 5 years ago - 2 comments

#99 - Why 'bias=not use_spectral_norm' ?

Issue - State: open - Opened by cmyyy over 5 years ago

#98 - No such file or directory

Issue - State: open - Opened by landycy over 5 years ago

#97 - C:\Users\86199>python D:\\edge-connect-master\\scripts\\metrics.py --data-path "D:\\edge-connect-master\\input" --output-path D:\\edge-connect-master\\output [data_path] = D:\\edge-connect-master\\input [output_path] = D:\\edge-connect-master\\output [debug] = 0 Traceback (most recent call last): File "D:\\edge-connect-master\\scripts\\metrics.py", line 48, in <module> img_pred = (imread(str(path_pred) + '\\'+basename(str(fn))) / 255.0).astype(np.float32) File "C:\Anaconda3\lib\site-packages\numpy\lib\utils.py", line 101, in newfunc return func(*args, **kwds) File "C:\Anaconda3\lib\site-packages\scipy\misc\pilutil.py", line 164, in imread im = Image.open(name) File "C:\Anaconda3\lib\site-packages\PIL\Image.py", line 2652, in open fp = builtins.open(filename, "rb") FileNotFoundError: [Errno 2] No such file or directory: 'D:\\\\edge-connect-master\\\\output\\places2_01.png'

Issue - State: open - Opened by landycy over 5 years ago

#96 - dataset_edges_train/val/test.flist

Issue - State: open - Opened by swhan0329 over 5 years ago - 5 comments

#95 - How to test in HED detection?

Issue - State: closed - Opened by GuardSkill over 5 years ago - 5 comments

#94 - style loss

Issue - State: closed - Opened by Turlan over 5 years ago

#93 - How to generate train, test, and validation set file lists

Issue - State: open - Opened by ghost over 5 years ago - 1 comment

#92 - Test error.

Issue - State: open - Opened by SunHu1 over 5 years ago - 5 comments

#91 - ImportError: cannot import name 'imread' from 'scipy.misc' (C:\Users\admin\AppData\Local\Programs\Python\Python37\lib\site-packages\scipy\misc\init.py)

Issue - State: open - Opened by Nakasu-ksm over 5 years ago - 2 comments

#90 - Runtime error when training joint model

Issue - State: open - Opened by ShnitzelKiller over 5 years ago - 7 comments

#89 - question about the test results

Issue - State: closed - Opened by SunHu1 over 5 years ago

#88 - KeyError: 'module.encoder.1.weight_orig'

Issue - State: open - Opened by luistelmocosta over 5 years ago - 3 comments
Labels: Bug

GitHub / knazeri/edge-connect issues and pull requests