cdahms123 / argo-3d-det-second-pointpillars

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

argo-3d-det-second-pointpillars

steps to run:

1) Download and Setup Argoverse Tracking

Make a directory ~/DownloadsArgoverse and from https://www.argoverse.org/data.html under Argoverse 1.1 download the following:

Under Simple Datasets v1.1:
3D Tracking link (filename tracking_sample_v1.1.tar.gz)

Under Argoverse HD Maps:
Miami and Pittsburgh link (filename hd_maps.tar.gz)

Under Argoverse 3D Tracking v1.1:
Training Part 1 link (filename tracking_train1_v1.1.tar.gz)
Training Part 2 link (filename tracking_train2_v1.1.tar.gz)
Training Part 3 link (filename tracking_train3_v1.1.tar.gz)
Training Part 4 link (filename tracking_train4_v1.1.tar.gz)
Validation link (filename tracking_val_v1.1.tar.gz)
Testing link (filename tracking_test_v1.1.tar.gz)

Make a directory ~/argoverse-tracking and extract/move/rearrange the above downloads into it so it ends up as follows:

~/argoverse-tracking
    /sample  (1 trip)
    /test    (24 trips)
    /train   (65 trips)
    /val     (24 trips)

Note the following:
-after extracting, the folders with the trip data are nested multiple levels
-the train directory has to be made and the trips from train1, train2, train3, and train4 have to be moved into train
-the hd_maps.tar.gz file (from the Argoverse HD Maps -> Miami and Pittsburgh link) is not used in this step, but rather is put inside the argoverse-api clone in the following steps.

2) Clone argoverse-api repository:

cd ~
git clone https://github.com/argoai/argoverse-api.git

3) Download and move map files into clone

Extract the hd_maps.tar.gz file (from the Argoverse HD Maps -> Miami and Pittsburgh link), this will provide the directory map_files. Move the extracted directory map_files inside the root of your argoverse-api clone. When done, your argoverse-api directory structure should look like this:

~/argoverse-api
    /argoverse/...
    /demo_usage/...
    /docs/...
    /images/...
    /integration_tests/...
    /map_files/...        <-- this directory has to be manually added per the instructions above, all others are simply part of the clone
    /sphinx/...
    /tests/...
    /Argoverse-Terms_of_Use.txt
    /CONTRIBUTING.md
    /LICENSE
    /MANIFEST.in
    /pyproject.toml
    /README.md
    /setup.cfg
    /setup.py
    /tox.ini
    

4) Install argoverse-api as a pip package

Note: The instructions for this step deviate from the https://github.com/argoai/argoverse-api readme instructions step 4, which I could not get to work. These instructions involve making a copy of the map data into the pip package install location which is sloppy, but it works.

ToDo: Reconcile official instructions to these eventually.

cd ~/argoverse-api
sudo python3 setup.py install
sudo cp -R /home/cdahms/argoverse-api/map_files /usr/local/lib/python3.8/dist-packages/argoverse-1.1.0-py3.8.egg/

ToDo: write rest of this when content is complete

999) Submit

Upload file to https://eval.ai/web/challenges/challenge-page/725/submission.

Check submission at https://eval.ai/web/challenges/challenge-page/725/my-submission.

When submission has processed, view leaderboard https://eval.ai/web/challenges/challenge-page/725/leaderboard/1974

About


Languages

Language:Python 100.0%