Pip: Move from tensorflow-gpu to tensorflow package (#1530)

* Pip: Move from tensorflow-gpu to tensorflow package

Since tensorflow 2.1 the regular tensorflow package contains GPU support and replaces the tensorflow-gpu package, which will be depreceated at some point.

* Readme: Update Tensorflow version

* Update Pipfile.lock

Co-authored-by: Willem Melching <willem.melching@gmail.com>
pull/1547/head
Ewout ter Hoeven 2020-05-20 22:17:54 +02:00 committed by GitHub
parent b1795235f9
commit 3f8afe593f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 2 additions and 2 deletions

View File

@ -89,7 +89,7 @@ pipenv install --system --deploy
pip install -r tools/requirements.txt
# to make modeld work on PC with nvidia GPU
pip install tensorflow-gpu==2.2
pip install tensorflow==2.2
# for loggerd to work on ubuntu
# TODO: PC should log somewhere else

View File

@ -17,7 +17,7 @@ git clone https://github.com/commaai/openpilot.git
```
- Follow [this readme](https://github.com/commaai/openpilot/tree/master/tools) to install the requirements
- Add line "export PYTHONPATH=$HOME/openpilot" to your ~/.bashrc
- You also need to install tensorflow-gpu 2.1.0 (if not working, try 2.0.0) and nvidia drivers: nvidia-xxx/cuda10.0/cudnn7.6.5
- You also need to install tensorflow 2.2 and nvidia drivers: nvidia-xxx/cuda10.0/cudnn7.6.5
- Install [OpenCL Driver](http://registrationcenter-download.intel.com/akdlm/irc_nas/12556/opencl_runtime_16.1.2_x64_rh_6.4.0.37.tgz)
- (Note: the code assumes cl platforms order to be 0.GPU/1.CPU when running clinfo; if reverse, change the -1 to -2 in selfdrive/modeld/modeld.cc#L130; helping us refactor this mess is encouraged)
- Install [OpenCV4](https://www.pyimagesearch.com/2018/08/15/how-to-install-opencv-4-on-ubuntu/) (ignore the Python part)