kujason / avod

Code for 3D object detection for autonomous driving
MIT License
931 stars 349 forks source link

run_inference problem #143

Open BigRayLee opened 5 years ago

BigRayLee commented 5 years ago

Hello,when I run inference on test dataset. And I tried to generate my own plane file. The parameters look like 0.0119063 -0.999904 0.00707348 1.63311. And the problem occurred when inference Step 120000: 276 / 7518, Inference on sample 000275. File "/home/bit201/lr_avod/avod-ssd-master/avod/core/bev_generators/bev_slices.py", line 94, in generate_bev height_map = np.zeros((voxel_grid_2d.num_divisions[0], UnboundLocalError: local variable 'voxel_grid_2d' referenced before assignment And I think it might be the plane parameters problem. So I changed all the parameters to 0.0e+00 -1.0e+00 0.0e+00 1.65e+00. But the problem still occurred in Step 120000: 5233 / 7518, Inference on sample 005232. So I want to ask what is the reason for this? And how to solve it?

yinhai86924 commented 4 years ago

Hello,when I run inference on test dataset. And I tried to generate my own plane file. The parameters look like 0.0119063 -0.999904 0.00707348 1.63311. And the problem occurred when inference Step 120000: 276 / 7518, Inference on sample 000275. File "/home/bit201/lr_avod/avod-ssd-master/avod/core/bev_generators/bev_slices.py", line 94, in generate_bev height_map = np.zeros((voxel_grid_2d.num_divisions[0], UnboundLocalError: local variable 'voxel_grid_2d' referenced before assignment And I think it might be the plane parameters problem. So I changed all the parameters to 0.0e+00 -1.0e+00 0.0e+00 1.65e+00. But the problem still occurred in Step 120000: 5233 / 7518, Inference on sample 005232. So I want to ask what is the reason for this? And how to solve it?

hello, how to run your test dataset, it need labels folder? I wan to konw your structure of test dataset , it similar to training data as readme said?

zhaotudou commented 4 years ago

helle, have you solved this problem yet? I also encountered this problem, thank you

BigRayLee commented 4 years ago

helle, have you solved this problem yet? I also encountered this problem, thank you

I don't solve this problem, but I changed the computer, this problem will not occur in some situations.

BigRayLee commented 4 years ago

Hello,when I run inference on test dataset. And I tried to generate my own plane file. The parameters look like 0.0119063 -0.999904 0.00707348 1.63311. And the problem occurred when inference Step 120000: 276 / 7518, Inference on sample 000275. File "/home/bit201/lr_avod/avod-ssd-master/avod/core/bev_generators/bev_slices.py", line 94, in generate_bev height_map = np.zeros((voxel_grid_2d.num_divisions[0], UnboundLocalError: local variable 'voxel_grid_2d' referenced before assignment And I think it might be the plane parameters problem. So I changed all the parameters to 0.0e+00 -1.0e+00 0.0e+00 1.65e+00. But the problem still occurred in Step 120000: 5233 / 7518, Inference on sample 005232. So I want to ask what is the reason for this? And how to solve it?

hello, how to run your test dataset, it need labels folder? I wan to konw your structure of test dataset , it similar to training data as readme said?

sorry for late reply, it doesn't need the label folder, the data structure is similar to the training data.

zhaotudou commented 4 years ago

Thank you for your reply. It's very exciting to get your reply soon.

------------------ 原始邮件 ------------------ 发件人: "BigRayLee"<notifications@github.com>; 发送时间: 2020年5月21日(星期四) 下午5:28 收件人: "kujason/avod"<avod@noreply.github.com>; 抄送: "甜土豆"<1292408887@qq.com>;"Comment"<comment@noreply.github.com>; 主题: Re: [kujason/avod] run_inference problem (#143)

I don't solve this problem, but I changed the computer, this problem will not occur in some situations.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

Pari-singh commented 4 years ago

I believe this is the issue when the bev_slices are generating slice points <=1. In that case, the initialization of voxel_grid2d, which is happening inside the if condition is getting skipped and hence when accessed outside the if condition, is throwing the error.

I am currently struggling with the same issue, trying to figure out why there are no slice points being created, I am running inference on Kitti but in the camera space.

kkk324 commented 3 years ago

@Pari-singh Did you solve this issue?

bm-93 commented 3 years ago

did anyone of you solve this issue? I am facing the same problem. any help is appreciated.

qiaoli96 commented 2 years ago

Sorry,did anyone of you solve this issue? Thank you for your reply.

duuxu commented 2 years ago

您好,当我在测试数据集上运行推理时。我试图生成我自己的平面文件。参数类似于 。并且当推理步骤120000:276 / 7518,对样本000275进行推理时出现问题。 我认为这可能是平面参数问题。所以我把所有的参数都改成了.但问题仍然发生在 .所以我想问一下,这是什么原因?如何解决它?0.0119063 -0.999904 0.00707348 1.63311``File "/home/bit201/lr_avod/avod-ssd-master/avod/core/bev_generators/bev_slices.py", line 94, in generate_bev height_map = np.zeros((voxel_grid_2d.num_divisions[0], UnboundLocalError: local variable 'voxel_grid_2d' referenced before assignment ``0.0e+00 -1.0e+00 0.0e+00 1.65e+00``Step 120000: 5233 / 7518, Inference on sample 005232

Hello, I also encountered this problem, how did you solve it? Can you share your method of requesting planes data?

qiaoli96 commented 2 years ago

我已收到您的邮件,谢谢!--李悄(^_^)

snehahegdek commented 4 months ago

Hello, did anyone solve the issue.

qiaoli96 commented 4 months ago

我已收到您的邮件,谢谢!--李悄(^_^)