Release testing for Donkeycar 5.0.dev3 #1142
Comments
simulatorI can confirm all tests passed for the simulator using DC 5.0dev3 running on Intel Linux (Ubuntu 22.04) using the latest gym-donkeycar & simulator release (22.11.06 from Nov. 6th, 2022) necessary adjustments in gym-donkeycar: use gym==0.22 instead of version 0.21 in setup.py |
@Heavy02011 - can you submit this as a PR in gym-donkeycar? I'm sure @Maximellerbach can test and merge this. |
|
done |
conda on Jetson Nano Orin 8GB (deprecated)
Can someone confirm this? |
venv on Jetson Nano Orin 8GB
|
|
Failed to run donkey ui on Windows WSL Ubuntu 22.04. Kivy is version 2.1.0 This is how the install went; |
Pi 3B, 1GB: OS Lite Bullseye (64-bit)
|
RPi3B, 1GB: OS Bullseye (64-bit)
|
@Ezward - can you please see if |
@Heavy02011 - if the above install gets you opencv, then you are running out of a |
|
@DocGarbanzo, I think that we saw this error when installing DC 4.4.X on JP5.0.2 when using conda? |
|
try
try to remove kivy-jetson from install und try to use kivy==2.1.0 |
I followed exactly the docs and did not get an opencv installed, thats why I added it and it worked |
Ok cool, thanks @Heavy02011 - I'll try to reproduce (using RPi 4, but this shouldn't make a difference). |
|
@autorope/contributors - I updated the install instructions above with the recipe of going back to a |
|
@DocGarbanzo, TCIII |
|
@DocGarbanzo I attempted to install on my MacBook Pro with M2-Max, MacOS Ventura 13.4.1 (c).
|
@Ezward - for zsh you need to backslashify the |
|
Good news. All functions tested on WSL and working! |
|
@DocGarbanzo I retried with backslashes. This time it ran but I got a failure: |
Hi @Ezward Please just try again with the new target |
|
@DocGarbanzo It failed on Mac M2 with a different message: |
@Ezward - if you are using pip install --extra-index-url https://test.pypi.org/simple/ donkeycar_pypi\[macos\] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Below is a test matrix
Install Donkey Car software with pip
Unfortunately we are still tied to the system python dependencies on both robot platforms, RPi and Jetson. Therefore we can only work in an isolated
condaenvironment on the PC and have to usevenvon the robot platforms.On the PC
The new installation with pip requires creating a
condaenvironment with python 3.9 first and then installing donkeycar into this environment. We are using that same install on all operating systems.where
XXstands forpcif you are on Intel Mac or Linux andmacosif you are on Apple Arm. The changes for the official do are on the branch here: Install Software on Host PCOn the Pi or Nano
The installation on Jetson using JP 5.X will create a python 3.8 env. The reason is that NVidia only provides for a python 3.8 / tensorflow 2.9 package. On the RPi w/ bullseye we will create a python 3.9 env instead.
where
XXstands forpiornano.Testing
On each of the platforms listed below we want to
Install Software on Donkeycar
On each of the platforms listed below we want to
completetemplatepath_followtemplatecv_controltemplateWhen running the autopilots we just want to know that they start without crashing, so it is not important to actually drive a course.
The text was updated successfully, but these errors were encountered: