My submission ran on the server for a long time(about 20 mins)，then encountered a error : Your submission timed out after 00:15:00 minutes. What's wrong with this ,can anyone help me，thanks！
", "messages": [ { "id": "319423", "postDate": "04/26/2018 02:52:52", "content": "My submission ran on the server for a long time(about 20 mins)，then encountered a error : Your submission timed out after 00:15:00 minutes. What's wrong with this ,can anyone help me，thanks！
", "rawMarkdown": "My submission ran on the server for a long time(about 20 mins)，then encountered a error : Your submission timed out after 00:15:00 minutes. What's wrong with this ,can anyone help me，thanks！", "votes": null }, { "id": "320244", "postDate": "04/28/2018 01:50:13", "content": "Mine too.
", "rawMarkdown": "Mine too.", "votes": null }, { "id": "320245", "postDate": "04/28/2018 01:53:11", "content": "You have too many predictions.\nIf you have N predictions and that frame has M ground truth, the complexity for evaluation will be O(NM). If N is large than simply evaluation will yake too much time and time out
", "rawMarkdown": "You have too many predictions.\nIf you have N predictions and that frame has M ground truth, the complexity for evaluation will be O(NM). If N is large than simply evaluation will yake too much time and time out", "votes": null }, { "id": "320711", "postDate": "04/29/2018 15:12:18", "content": "My submission of ~11,000 lines took >12min. and one with >20,000 lines timed out. Training data has >11 objects per image. If test data follow the same, the output would contain >20,000 lines. Could the organizer please enlarge the timeout threshold please?
", "rawMarkdown": "My submission of ~11,000 lines took >12min. and one with >20,000 lines timed out. Training data has >11 objects per image. If test data follow the same, the output would contain >20,000 lines. Could the organizer please enlarge the timeout threshold please?", "votes": null }, { "id": "321187", "postDate": "04/30/2018 19:27:26", "content": "We will make the submission evaluation time limit 20 mins for this competition. Please note that it posts significant load on our scoring server, so we will be monitoring closely, and might have to reduce the number of allowed submissions to 1/day if our server can't handle it.
", "rawMarkdown": "We will make the submission evaluation time limit 20 mins for this competition. Please note that it posts significant load on our scoring server, so we will be monitoring closely, and might have to reduce the number of allowed submissions to 1/day if our server can't handle it.", "votes": null }, { "id": "321273", "postDate": "04/30/2018 22:52:06", "content": "You change of timing threshold definitely helps in my case. Thanks! The submission I intended to submit went through and the score is slightly higher (0.03 vs 0.02). I'll really appreciative it if you could also confirm row order to be the same as OpenCV, which is what I'm currently using. Or ideally, it would be really helpful if you could release the encoding of a non-trivial label image (it could be a training image) as an example. Your sample submission is rather confusing:
\n\n\n001bbdb5c4f43bd4dc2b3e3a08b7202a,33,300,1,0 100|0 100|0 100|\n
The three parts of the encoding all starts from pixel 0. So there are really only 100 pixels, while the total pixel counting is 300.
\n\nAm I missing something really big?
", "rawMarkdown": "You change of timing threshold definitely helps in my case. Thanks! The submission I intended to submit went through and the score is slightly higher (0.03 vs 0.02). I'll really appreciative it if you could also confirm row order to be the same as OpenCV, which is what I'm currently using. Or ideally, it would be really helpful if you could release the encoding of a non-trivial label image (it could be a training image) as an example. Your sample submission is rather confusing:\n\n```\n001bbdb5c4f43bd4dc2b3e3a08b7202a,33,300,1,0 100|0 100|0 100|\n```\n\nThe three parts of the encoding all starts from pixel 0. So there are really only 100 pixels, while the total pixel counting is 300.\n\nAm I missing something really big?", "votes": null }, { "id": "321274", "postDate": "04/30/2018 23:00:01", "content": "@Wei Dong, have you checked out convertVideotoCSV.py?
Yes. My understanding is that it's using row-major/OpenCV order. See my last reply here:\nhttps://www.kaggle.com/c/cvpr-2018-autonomous-driving/discussion/54573
\n\nI'll try to run it.
\n\nUpdate:\nI ran the code. I can confirm the provided code uses row major/C order and is the SAME as opencv.
\nEverything is OK except that \"which is different from OpenCV.\" on evaluation page is disturbing.
Mine too.
", "votes": null, "replies": [] }, { "id": 320245, "author_name": "woodywang", "author_url": "", "post_date": "04/28/2018 01:53:11", "content": "You have too many predictions.\nIf you have N predictions and that frame has M ground truth, the complexity for evaluation will be O(NM). If N is large than simply evaluation will yake too much time and time out
", "votes": null, "replies": [] }, { "id": 320711, "author_name": "aaalgo", "author_url": "", "post_date": "04/29/2018 15:12:18", "content": "My submission of ~11,000 lines took >12min. and one with >20,000 lines timed out. Training data has >11 objects per image. If test data follow the same, the output would contain >20,000 lines. Could the organizer please enlarge the timeout threshold please?
", "votes": null, "replies": [] }, { "id": 321187, "author_name": "wendykan", "author_url": "", "post_date": "04/30/2018 19:27:26", "content": "We will make the submission evaluation time limit 20 mins for this competition. Please note that it posts significant load on our scoring server, so we will be monitoring closely, and might have to reduce the number of allowed submissions to 1/day if our server can't handle it.
", "votes": null, "replies": [] }, { "id": 321273, "author_name": "aaalgo", "author_url": "", "post_date": "04/30/2018 22:52:06", "content": "You change of timing threshold definitely helps in my case. Thanks! The submission I intended to submit went through and the score is slightly higher (0.03 vs 0.02). I'll really appreciative it if you could also confirm row order to be the same as OpenCV, which is what I'm currently using. Or ideally, it would be really helpful if you could release the encoding of a non-trivial label image (it could be a training image) as an example. Your sample submission is rather confusing:
\n\n\n001bbdb5c4f43bd4dc2b3e3a08b7202a,33,300,1,0 100|0 100|0 100|\n
The three parts of the encoding all starts from pixel 0. So there are really only 100 pixels, while the total pixel counting is 300.
\n\nAm I missing something really big?
", "votes": null, "replies": [ { "id": 321274, "author_name": "wendykan", "author_url": "", "post_date": "04/30/2018 23:00:01", "content": "@Wei Dong, have you checked out convertVideotoCSV.py?
Yes. My understanding is that it's using row-major/OpenCV order. See my last reply here:\nhttps://www.kaggle.com/c/cvpr-2018-autonomous-driving/discussion/54573
\n\nI'll try to run it.
\n\nUpdate:\nI ran the code. I can confirm the provided code uses row major/C order and is the SAME as opencv.
\nEverything is OK except that \"which is different from OpenCV.\" on evaluation page is disturbing.