Closed HRHLALALA closed 2 months ago
@Eddie-JUB @JunyongYun-SPA I am trying to generate the preprocessed dataset for YNet but failed. I have run into the following problems:
./SiT_dataset/first_loc
at https://github.com/SPALaboratory/SiT-Dataset/blob/1409c5cd98ab872a5a7223ee7b5abed17173aa05/prediction/ynet/preprocess_ped.py#L205. label_3d
and draw the polygons in the json files in the folder semantic-maps
. However, they are not matched as well. label_3d
.Would you please provide some details of the data format and data preprocessing? Really appreciate in advance.
sorry to be late abour updaring code. i will upload as soon as possible for ped prediction.
Jungho, Kim Ph.D. Candidate, Dept. of Artificial Intelligence, Hanyang University.
Signal Processing & Artificial-intelligence Laboratory. Hanyang University Engineering Center Main Building 408-1 222 Wangsimni-ro, Seongdong-gu, Seoul 133791, Republic of Korea
E-mail : @. @.>
Mobile : +82 10-9852-6898
2024년 8월 2일 (금) 18:08, HRHLALALA @.***>님이 작성:
I am trying to generate the preprocessed dataset for YNet but failed. I have run into the following problems:
- I cannot find the ./SiT_dataset/first_loc at https://github.com/SPALaboratory/SiT-Dataset/blob/1409c5cd98ab872a5a7223ee7b5abed17173aa05/prediction/ynet/preprocess_ped.py#L205 .
- Even I comment the code and successfully generate the pickle files, the coordinates do not match the segmentation maps.
- I also tried to plot the locations in label_3d and draw the polygons in the json files in the folder semantic-maps. However, they are not matched as well.
Would you please provide some details of the data format and data preprocessing? Really appreciate in advance.
— Reply to this email directly, view it on GitHub https://github.com/SPALaboratory/SiT-Dataset/issues/5#issuecomment-2264924064, or unsubscribe https://github.com/notifications/unsubscribe-auth/BAFKBS4GBZMMSY7C2TCWVITZPNECNAVCNFSM6AAAAABLV4QTAWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENRUHEZDIMBWGQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>
I have the same problem. I would appreciate it if you could share the ynet config file as well.
We apologize for the delay in responding. We needed some extra time to review and verify the code. Could you please check the newly uploaded Ynet code?
We will be re-uploading the NSP code shortly as well.
Thank you for your understanding.
We apologize for the delay in responding. We needed some extra time to review and verify the code. Could you please check the newly uploaded Ynet code?
We will be re-uploading the NSP code shortly as well.
Thank you for your understanding.
@SPA-junghokim Thank you for your update. I notice that Cafeteria_4 also does have the label_3d folder. Can you help check this? Thanks
The reason why the label_3d is missing for Cafeteria_4 is that this scene is part of the test set. Similarly, other scenes that belong to the test set do not have label_3d either.
The reason why the label_3d is missing for Cafeteria_4 is that this scene is part of the test set. Similarly, other scenes that belong to the test set do not have label_3d either.
@SPALaboratory One main issue is that this could cause error when executing preprocess_ped.py. Removing the test sets successfull generate the pandas files. The trajectories match the scene map very well. Thank you very much.
BTW, when will the trajectories of the test set release?
Thanks for checking the code and letting me know. Thanks to you, I've fixed the code.
Regarding the test set, there are still some things to review, so it will take some more time.
Hi, the full dataset is too large to be downloaded. Is it possible to release the preprocessed trajectory files such as ``trajectory_txt", train.pkl, val.pkl and test.pkl?
ynet ├── data │ ├── semantic_maps_json │ ├── semantic_map │ ├── trajectory_txt │ ├── tran.pkl │ ├── val.pkl │ ├── test.pkl