Bug. No such file or directory: './data//train.txt'

[GabbySuwichaya]

Thank you for sharing the repo! and for producing this awesome CVPR paper !!!

It is very interesting to use unsupervised learning with GAN in extracting the dominant plane between image pairs.

Here, I would like to learn from your code. However, as I try to run I have run evaluate.py and set model_dir and restore_file as in the following settings. I run into a bug. Therefore, I would like to kindly ask for your help.

• Could you please kindly help with the following bug? It seems that data/train.txt is missing....

I have run evaluate.py and set model_dir and restore_file as follows:

parser.add_argument('--model_dir', default='experiments/fine_tuning/', help="Directory containing params.json")
parser.add_argument('--restore_file', default='experiments/fine_tuning/fine_tuning.pth')

• Also do you have any demo to see how to use the homography estimation?

[MingboHong]

Hi, Thank you for your interest~~ You need to prepare the Deep Homography Dataset (for more detail can be found at readme.md) or you can customize your test dataset.

1. mkdir "data"
2. put the images in "data/images", and images list (such as "img_pair_1.jpg img_pair_2.jpg /n" format) in "data/test.txt"

[CarlosAlbertoToroArcila]

Good morning,

I did the database preprocessing as indicated in https://github.com/JirongZhang/DeepHomography and I also have the same problem FileNotFoundError: [Errno 2] No such file or directory: './data//train.txt'. What I did was to process the database with Run "video2img.py", is there any additional process I should do?, thank you very much for your help.

Best regards, Carlos A. Toro-Arcila

[lyh200095]

Is this problem solved?

[zhongchangxaing]

May I ask where you got video2img.py from?

[lyh200095]

You can get the file from the previous link of dataset in the GitHub of his team’s work content-aware homography estimation.

发自我的iPhone

------------------ Original ------------------ From: zhongchangxaing @.> Date: Fri,Dec 8,2023 10:59 AM To: megvii-research/HomoGAN @.> Cc: lyh @.>, Comment @.> Subject: Re: [megvii-research/HomoGAN] Bug. No such file or directory:'./data//train.txt' (Issue #3)

May I ask where you got video2img.py from?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[zhongchangxaing]

Hello, thanks for your answer, I have found the video2img.py file. However, after passing video2img.py, the generated image set is not paired. How can I generate a data set that can be directly trained or tested?

[lyh200095]

You can contact me on wechat 18203220989.I can send the txt to you.  

那一年,雪花飘落 @.***

 

------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2023年12月14日(星期四) 下午4:23 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [megvii-research/HomoGAN] Bug. No such file or directory: './data//train.txt' (Issue #3)

Hello, thanks for your answer, I have found the video2img.py file. However, after passing video2img.py, the generated image set is not paired. How can I generate a data set that can be directly trained or tested?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>