Can you show me your dataset structure？ about delg

[berooo]

The structure is:
DELG-master/datasets/data/GLDv2

GLDv2 is the dir of GLDv2-clean

[vodka95]

Thanks! Got it！But the GLDv2 dataset is too large for my server, can i use a smaller one such as 'oxford building' to replace it?

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "feymanpriv/DELG" <[email protected]&gt;; 发送时间:&nbsp;2020年12月29日(星期二) 下午2:50 收件人:&nbsp;"feymanpriv/DELG"<[email protected]&gt;; 抄送:&nbsp;" 蘼芜"<[email protected]&gt;;"Author"<[email protected]&gt;; 主题:&nbsp;Re: [feymanpriv/DELG] Can you show me your dataset structure？ (#3) The structure is: DELG-master/datasets/data/GLDv2 GLDv2 is the dir of GLDv2-clean — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[berooo]

Oxford building and oxford building are more suitable for testing but not for training. But if you have to use it as a training dataset, you may need to modify some code in this project.发自我的华为手机-------- 原始邮件 --------发件人： vodka95 <[email protected]>日期： 2020年12月29日周二 傍晚5:10收件人： feymanpriv/DELG <[email protected]>抄送： berooo <[email protected]>, Assign <[email protected]>主 题： Re: [feymanpriv/DELG] Can you show me your dataset structure？ (#3) Thanks! Got it！But the GLDv2 dataset is too large for my server, can i use a smaller one such as 'oxford building' to replace it?

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "feymanpriv/DELG" <[email protected]&gt;; 发送时间:&nbsp;2020年12月29日(星期二) 下午2:50 收件人:&nbsp;"feymanpriv/DELG"<[email protected]&gt;; 抄送:&nbsp;" 蘼芜"<[email protected]&gt;;"Author"<[email protected]&gt;; 主题:&nbsp;Re: [feymanpriv/DELG] Can you show me your dataset structure？ (#3) The structure is: DELG-master/datasets/data/GLDv2 GLDv2 is the dir of GLDv2-clean — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe. —You are receiving this because you were assigned.Reply to this email directly, view it on GitHub, or unsubscribe.

[vodka95]

&nbsp;Hi, When i try to use Oxford building and paris building dataset to test the model, it seems that i need three files like&nbsp; GLOBAL_FEATURE_PATH='roxford5k512gem_delg_res50_3global.mat' LOCAL_FEATURE_PATH='roxford5k__s512_localfea.pickle' ASMK_SCORE_PATH='localfeatures/delg_asmk_6553.pkl' then i turn to&nbsp;https://github.com/filipradenovic/revisitop&nbsp;to run the model. If i may be so bold, could you send me the three files? I’d be grateful if you could help me. Thanks!

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "feymanpriv/DELG" <[email protected]&gt;; 发送时间:&nbsp;2020年12月29日(星期二) 晚上6:38 收件人:&nbsp;"feymanpriv/DELG"<[email protected]&gt;; 抄送:&nbsp;" 蘼芜"<[email protected]&gt;;"Author"<[email protected]&gt;; 主题:&nbsp;Re: [feymanpriv/DELG] Can you show me your dataset structure？ (#3) Oxford building and oxford building are more suitable for testing but not for training. But if you have to use it as a training dataset, you may need to modify some code in this project.发自我的华为手机-------- 原始邮件 --------发件人： vodka95 <[email protected]&gt;日期： 2020年12月29日周二 傍晚5:10收件人： feymanpriv/DELG <[email protected]&gt;抄送： berooo <[email protected]&gt;, Assign <[email protected]&gt;主 题： Re: [feymanpriv/DELG] Can you show me your dataset structure？ (#3) Thanks! Got it！But the GLDv2 dataset is too large for my server, can i use a smaller one such as 'oxford building' to replace it?

------------------&amp;nbsp;原始邮件&amp;nbsp;------------------ 发件人: "feymanpriv/DELG" <[email protected]&amp;gt;; 发送时间:&amp;nbsp;2020年12月29日(星期二) 下午2:50 收件人:&amp;nbsp;"feymanpriv/DELG"<[email protected]&amp;gt;; 抄送:&amp;nbsp;" 蘼芜"<[email protected]&amp;gt;;"Author"<[email protected]&amp;gt;; 主题:&amp;nbsp;Re: [feymanpriv/DELG] Can you show me your dataset structure？ (#3) The structure is: DELG-master/datasets/data/GLDv2 GLDv2 is the dir of GLDv2-clean — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe. —You are receiving this because you were assigned.Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[superluckyxhh]

The structure is:
DELG-master/datasets/data/GLDv2

GLDv2 is the dir of GLDv2-clean

Hello.When I reviewed this repo. There is self._data_path, self._split = data_path, split in DataSet class and split is 'train_list.txt' in debuger. But Train dataset GLDv2 has not such as the 'txt' files. So I want to ask what wrong in my operations