Comments (6)
@fatemehazimi990 I see. Your guess is reasonable. In that case, how about you try either of the following ways?
- Add a
forecasting=True
in the config field ofvisualization
, just like the config field fortrain
- Name a new config pipeline called
visualization_pipeline_multiframe
like below, and replace thetrain_pipeline_multiframe
at with it. Essentially, the fields related to forecasting are removed.
visualization_pipeline_multiframe = [
dict(type='TrackResizeCropFlipImage', data_aug_conf = ida_aug_conf, training=True),
dict(type='TrackNormalizeMultiviewImage', **img_norm_cfg),
dict(type='TrackPadMultiViewImage', size_divisor=32),
dict(type='FormatBundle3DTrack'),
dict(type='Collect3D', keys=['points', 'gt_bboxes_3d', 'gt_labels_3d', 'instance_inds', 'img', 'timestamp', 'l2g_r_mat', 'l2g_t', 'l2g'])
]
I think the original config of loading gt forecasting was useful when I validated the behavior of my forecasting. But it caused some dirty part in the program.
from pf-track.
Awesome, the first option already solved the problem. Thanks ^^
from pf-track.
I tried generating the results for miniset using test_tracking.py, then running the visualization command for bev.py. I get the following error in raw_data = dataset[data_info_idx]
saying *** KeyError: 'gt_forecasting_locs
.
from pf-track.
@fatemehazimi990 Thanks for your question!
Based on my understanding, this error is not related to results.json
, but the dataset part. As I don't have the computation to run these experiments after finishing my internship (which is unfortunate), having the full commands and error messages woule help a lot in finding the issues. Hope it helps!
from pf-track.
@ziqipang Thanks for your quick reply. I use the run command python tools/video_demo/bev.py ./projects/configs/tracking/petr/f3_q500_800x320.py --result work_dir/f3_petr_800x320/results/mini/results_nusc_tracking.json --show-dir ./work_dirs/visualizations/
I believe it should be an issue with the config. There I see
visualization=dict(type=dataset_type, pipeline=train_pipeline,
pipeline_multiframe=train_pipeline_multiframe,
data_root=data_root, test_mode=False,
classes=class_names, modality=input_modality,
ann_file=data_root + 'tracking_forecasting-mini_infos_val.pkl',
num_frames_per_sample=1,)
so it uses the train_pipeline for visualization, and test_mode is False. At the same time, it doesnt seem train_pipline provides the expected keys ... Just my guesses :)
from pf-track.
btw the error is KeyError: 'gt_forecasting_locs'
which happens in this line
PF-Track/tools/video_demo/bev.py
Line 68 in a254761
from pf-track.
Related Issues (20)
- Question about Track Extension HOT 6
- About the motion prediction evaluation HOT 2
- Why two step training HOT 12
- question about distributed training HOT 6
- What is the result.json file in relation to visualization code? HOT 2
- about visualization HOT 2
- about result.json HOT 6
- Evaluaiton on the test set HOT 2
- for Validation HOT 6
- About detection performance HOT 4
- Docker file
- Didn't provide mmdet3d folder HOT 2
- Train and infer on KITTI data HOT 1
- Gradient checkpointing with ddp HOT 6
- About the tracking result HOT 5
- Get some Errors when test tracking method HOT 4
- Adjusting batch size HOT 4
- detection_head HOT 5
- nuscenes_dataset HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from pf-track.