PanopticBEV
PanopticBEV copied to clipboard
robot-learning-freiburg
Question about Model Training
Dear all, we trained the model on Nuscenes and Kitti360 respectively. The Nuscenes work fine. But for Kitti360, the results don't seem very well. We attached the result at the end. I didn't change any code or configuration. I found that the size of the FV image in Kitti360 is 768 * 1400 on paper but 384 * 1400 in the configuration file in code. Which one is correct?
01:16:57 ---------------- Semantic mIoU ---------------- 01:16:57 road : 0.66676 01:16:57 sidewalk : 0.29211 01:16:57 building : 0.26043 01:16:57 wall : 0.03577 01:16:57 vegetation : 0.34960 01:16:57 terrain : 0.14224 01:16:57 occlusion : 0.40774 01:16:57 person : 0.00000 01:16:57 rider : 0.00789 01:16:57 car : 0.35977 01:16:57 truck : 0.07531 01:16:57 ---------------- Panoptic Scores ---------------- 01:16:57 po_miou : 0.23538 01:16:57 sem_miou : 0.23615 01:16:57 pq : 0.14057 01:16:57 pq_stuff : 0.17719 01:16:57 pq_thing : 0.07647 01:16:57 sq : 0.55955 01:16:57 sq_stuff : 0.60958 01:16:57 sq_thing : 0.47198 01:16:57 rq : 0.21159 01:16:57 rq_stuff : 0.26759
Which data set of Nuscenes are you using?
How do you set these Settings?
I would appreciate it if you could give me a reply!
Which data set of Nuscenes are you using?
download each keyframe and combine them together
Thank you for the response! You are so kind! How Do you set these settings? I really need it!
---Original--- From: @.> Date: Fri, Mar 18, 2022 20:06 PM To: @.>; Cc: @.@.>; Subject: Re: [robot-learning-freiburg/PanopticBEV] Question about ModelTraining (Issue #10)
Which data set of Nuscenes are you using? How do you set these Settings? I would appreciate it if you could give me a reply!
download each keyframe and combine them together
— Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you commented.Message ID: @.***>
Thank you for the response! You are so kind! How Do you set these settings? I really need it! … ---Original--- From: @.> Date: Fri, Mar 18, 2022 20:06 PM To: @.>; Cc: @.@.>; Subject: Re: [robot-learning-freiburg/PanopticBEV] Question about ModelTraining (Issue #10) Which data set of Nuscenes are you using? How do you set these Settings? I would appreciate it if you could give me a reply! download each keyframe and combine them together — Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you commented.Message ID: @.***>
depend on your hardware. For me, I don't change anything except for deleting the master_ip_addr
I'm still a little confused. Could you please give me a screenshot of this code settings? Thank you very much.
---Original--- From: @.> Date: Fri, Mar 18, 2022 20:22 PM To: @.>; Cc: @.@.>; Subject: Re: [robot-learning-freiburg/PanopticBEV] Question about ModelTraining (Issue #10)
Thank you for the response! You are so kind! How Do you set these settings? I really need it! … ---Original--- From: @.> Date: Fri, Mar 18, 2022 20:06 PM To: @.>; Cc: @.@.>; Subject: Re: [robot-learning-freiburg/PanopticBEV] Question about ModelTraining (Issue #10) Which data set of Nuscenes are you using? How do you set these Settings? I would appreciate it if you could give me a reply! download each keyframe and combine them together — Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you commented.Message ID: @.***>
depend on your hardware. For me, I don't change anything except for deleting the master_ip_addr
— Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you commented.Message ID: @.***>
Hi @anslt,
Thanks for bringing it to our attention. The dimensions in the configuration file are correct, i.e., the input image for KITTI-360 is of size (1408 x 384).
Hi @anslt,
Thanks for bringing it to our attention. The dimensions in the configuration file are correct, i.e., the input image for KITTI-360 is of size (1408 x 384).
Hello @nikhilgosala , i would like to ask which pytorch and cuda version do you use? I used pytorch 1.9.0+cu111, and the latest inplace-abn. But there always exists cuda issue when i use multiple gpus to train.
@laozheng1 The --resume should point to the model_best.pth file that you obtain after training the model.
The --dataset_root_dir parameter should point to the root directory of the nuScenes dataset. The root directory of the nuScenes dataset contains folders such as samples, v1.0-trainval and so on.
@anslt We tested it using Pytorch 1.8.1, CUDA 11.1 and the latest inplace-abn. We have had no issues with CUDA on our end.
@anslt We tested it using Pytorch 1.8.1, CUDA 11.1 and the latest inplace-abn. We have had no issues with CUDA on our end.
@nikhilgosala Thank you very much! I am sorry that I still have issue on multiple GPU training. I also would like to ask the CUDNN version and which compiler you use when you compile the BEV
Which data set of Nuscenes are you using?
@laozheng1 please ask your file configuration successfully? could you send out the picture of your configuration? thank
Which data set of Nuscenes are you using?
download each keyframe and combine them together
@anslt hello, i just want to test and also need to download so the keyframes well.
@anslt We tested it using Pytorch 1.8.1, CUDA 11.1 and the latest inplace-abn. We have had no issues with CUDA on our end.
@nikhilgosala Did you test download all the keyframe data?