That final moment when he tries to "import tensorflow" was amazing. I know exactly how he feels. Great vide. thanks I was searching forever how to get this done, and ALL the resources where filled with random unique situations that had nothing to do with my issues!
Hi Sam, i am experiencing an error with numpy for wheels. can you confirm which numpy version you are using? I am assuming also that python version is 3.7. Also you seemed to omit installing numpy in the virtual environment, are you using numpy 1.21.6? kindly confirm. thanks so much
Thanks a lot for helping me to install tensorflow in raspberry pi for more than 2 weeks. Now after seeing your video i done it within 30 mins ... Thanks for uploading this valuable video..🙏🙏🙏
Seriously awesome. I struggled for ever to get keras to work on ipython in my Raspberry. I followed your instructions and used Tensorflow 2.3 and it works great. Thanks so much for this. I would suggest adding a ETH or BTC address for those who want to tip in your profile.
Dude, you rock. Just installed Tensorflow version 2.3.0, and now I have a new virtual environment "env" on my Desktop (in sub /tf_pi). Really, thank you so much for the video, you earned a new sub. Happy coding! :o
I am getting an error at the time of installing tensorflow after the wget command, what I get is: tensorflow-2.0.0-cp37-none-linux_armv7l.whl is not supported wheel on this platform. Is it because my python in the virtual env is 3.6.6?
Hi there, how can I add tensorflow-text. Since when I try to install it it say no matching distribution found for tensorflow-text sys_platform != Win32. Pip version: 21.1.2 Python: 3.7.3 Tensorflow: 2.3.0 I installed tensorflow by your method it worked but tensorflow text I don't know how to do that. Can you help me out?
Sam, thank you! I have been trying to get tensorflow installed/working on my pi zero W for days. Your method seems to work. Now I'm off to get my model built and tested. Thanks again for sharing this!
I like your content and your channel, but the question. Why would you choose to install TensorFlow on a raspberry Pi? And not a device that has an NPU? Forgive me if I misunderstand what TensorFlow is for but isn't that for machine learning?
It is for machine learning, but some applications don't require fancy machinery. If you are training small models or deploying pre-trained models, the Raspberry Pi works great!
i installed tensorflow==1.14,but when i import tensorflow ,it will output an error says that "E tensorflow/core/platform/hadoop/hadoop_file_system.cc:132] HadoopFileSystem load error: libhdfs.so" if i ignore it .tensorfow seems can be used normally. how can i fix it . i noticed that you have gaven an existed soultion that to build a virtual env.but my project can not use virtual env. do you have another soultion?
Please help i run ur commands but when i run the install command it's didn't work so please some help here the error (tensorflow.....it not supported wheel on this platform )
Hi Sam! I was installing TF on the virtual env on Pi. When I execute the last command, it says ' Could not install package due to EnvironmentError: No such file or directory '. Can you please help me with this?
Thank you! but when installing the packages, why did you use sudo on "sudo apt-get install -y openmpi-bin libopenmpi-dev " and "sudo apt-get install -y libatlas-base-dev", my understanding was that when using sudo it installs outside of the environment so are these packages meant to be out of the enviroment? or does it just work differently with apt-get?
I had to read up on this. I'm pretty sure the sudo command is required for a direct download using apt-get. If I didn't want to use sudo, I could compile the package from source, but that is a few extra commands. From what I understand, apt-get is a tool to install OS related packages, and virtualenv is a tool to isolate Python environments. So a virtualenv will never be able to isolate packages installed using apt-get. If you're looking to contain packages downloaded using apt-get, I think "chroot" would be the direction to go. (I haven't used this before but SO told me so.)
ERROR: tensorflow-2.0.0-cp37-none-linux_armv7l.whl is not a supported wheel on this platform. I followed the instructions but still got this error after several attempts. May I know how to solve it?
I made a 2022 update for this tutorial last week. There are a couple differences so that could be a good place to start ruclips.net/video/QLZWQlg-Pk0/видео.html
In a virtual environment the --user is not applicable as it doesn't require any specific installation configuration as it's installed within the environment
Hello There, First of all, thanks for your video! Secondly, could you pls tell me how to load the virtual environment in Mu or any other code editor? Thanks Sandor
while training my model I'm facing the following error on my terminal! 2020-03-31 19:39:53.793083: W tensorflow/core/framework/cpu_allocator_impl.cc:81] Allocation of 188160000 exceeds 10% of system memory. Can you please suggest me how to increase my virtual environment memory?
@@SamWestby I'd like to know deeper about this problem if you can help me, like if the RAM is responsible or the micro SD card. And if the latter is the issue if I can extend the storage with any USB drive
im kind of new to this, using a pi 3 when I install tensorflow it returns that it is not a supported wheel on this platform any idea what of what I doing wrong?
@@SamWestby My issue is to learn tensoflow on raspberry Pi 4 as I don't have a windows PC. I am also very eager to learn the concepts of tensoflow also. Also I am keen to know how to make/train a training model. Thanks
@@SamWestby I did check and found that I need a full tensoflow version (not lite) to create a model then I can transfer the model to tensoflow lite format. I do require a demo like video where I can create a training model in full tensoflow. So in a sense I am facing touble training models. Apart from that can I update the tensoflow 2.0.0, which you showed in this video to 2.1.0 (the current version) in my Raspberry Pi 4? I have the 4 GB model.
@@SamWestby while training my model I'm facing the following error on my terminal! 2020-03-31 19:39:53.793083: W tensorflow/core/framework/cpu_allocator_impl.cc:81] Allocation of 188160000 exceeds 10% of system memory. I was following the official tutorial at tensorflow.org as you guided. The problem rose while I was making the def epoch programming. I checked over 10 times but the same prompt was shown to me every time. Can you please guide me through this? Thanks.
Try installing virtualenv with pip by running "python3 -m pip install virtualenv" If it says something along the lines of "installed virtualenv successfully" then "virtualenv env" should work now
@@SamWestby Okay I did that but now I am getting this error: ERROR: tensorflow-2.0.0-cp37-none-linux_armv7l.whl is not a supported wheel on this platform.
It is sad that Tensorflow do not have clear instruction on installing Tensorflow . Thanks for your help.
That final moment when he tries to "import tensorflow" was amazing. I know exactly how he feels. Great vide. thanks I was searching forever how to get this done, and ALL the resources where filled with random unique situations that had nothing to do with my issues!
Thank you! That's exactly why I made this video. It took me wayyyyy to long to figure this out
Hi Sam, i am experiencing an error with numpy for wheels. can you confirm which numpy version you are using? I am assuming also that python version is 3.7. Also you seemed to omit installing numpy in the virtual environment, are you using numpy 1.21.6? kindly confirm. thanks so much
Thanks a lot for helping me to install tensorflow in raspberry pi for more than 2 weeks. Now after seeing your video i done it within 30 mins ... Thanks for uploading this valuable video..🙏🙏🙏
Happy to help!
Seriously awesome. I struggled for ever to get keras to work on ipython in my Raspberry. I followed your instructions and used Tensorflow 2.3 and it works great. Thanks so much for this. I would suggest adding a ETH or BTC address for those who want to tip in your profile.
Thanks Tom! I'm happy I could help :)
Dude, you rock. Just installed Tensorflow version 2.3.0, and now I have a new virtual environment "env" on my Desktop (in sub /tf_pi). Really, thank you so much for the video, you earned a new sub. Happy coding! :o
Thanks! Let me know if there are any other videos you'd like me to make
I can't install tensorflow on my rpi 4 having 64 bit os.. Just got it freshly os installed.. How to proceed.. Any idea please
Thank you so very much my guy, I am new to Pi's and have been stuck for hours trying to get this simple stage completed
literally I had the same look on ur face, when I finally got it to install. thx u sm
I am getting an error at the time of installing tensorflow after the wget command, what I get is: tensorflow-2.0.0-cp37-none-linux_armv7l.whl is not supported wheel on this platform. Is it because my python in the virtual env is 3.6.6?
Yes you need to use Python 3.7
Thank you!! Now I have installed tensorflow 2.4.0 on my raspberry 2. Perfect tutorial, congratulations
Glad it helped!
There is now version v2.1.0, which was released a few days ago (March 17, 2020).
Just change the 2 reference numbers of 2.0.0 to 2.1.0 for the wget.
Thanks for the update!
Hi there, how can I add tensorflow-text. Since when I try to install it it say no matching distribution found for tensorflow-text sys_platform != Win32.
Pip version: 21.1.2
Python: 3.7.3
Tensorflow: 2.3.0
I installed tensorflow by your method it worked but tensorflow text I don't know how to do that. Can you help me out?
Sam, thank you! I have been trying to get tensorflow installed/working on my pi zero W for days. Your method seems to work. Now I'm off to get my model built and tested. Thanks again for sharing this!
Glad it helped!
I like your content and your channel, but the question. Why would you choose to install TensorFlow on a raspberry Pi? And not a device that has an NPU? Forgive me if I misunderstand what TensorFlow is for but isn't that for machine learning?
It is for machine learning, but some applications don't require fancy machinery. If you are training small models or deploying pre-trained models, the Raspberry Pi works great!
@@SamWestby interesting. Thank you for responding. It looks like it might be worth taking a small course in machine learning.
@@BrianThomas Definitely!
Got 2.4.0 to work on my RPi 4 from this tutorial. Thanks!
Thanks so much Sam. I have been trying to install Tensorflow on my rip 4 for so long. It worked.
Please does anyone know how I can install keras
i installed tensorflow==1.14,but when i import tensorflow ,it will output an error says that "E tensorflow/core/platform/hadoop/hadoop_file_system.cc:132] HadoopFileSystem load error: libhdfs.so"
if i ignore it .tensorfow seems can be used normally. how can i fix it . i noticed that you have gaven an existed soultion that to build a virtual env.but my project can not use virtual env. do you have another soultion?
Please help i run ur commands but when i run the install command it's didn't work so please some help here the error (tensorflow.....it not supported wheel on this platform )
I didn't create a virtual env before installing TensorFlow dependencies, should I uninstall it? and if yes how?
Thanks!!!!!!!!
I just installed the 2.3 version following your instruction :)
I'm happy I could help :)
Hi Sam! I was installing TF on the virtual env on Pi. When I execute the last command, it says ' Could not install package due to EnvironmentError: No such file or directory '. Can you please help me with this?
Question, if its in a virtual environment how can you make a program work if the tensor is in a seperate environment?
i get this error :ERROR: tensorflow-2.2.0-cp37-none-linux_armv7l.whl is not a supported wheel on this platform.
Hi ! Thanks for this ! But doing "virtualenv env", i have the error "command not found", does someone understand that ?
Try a Google search for "how to create a virtualenv python"
My pi cannot find the command virtualenv, is there perhaps a new version ?
Thank you! but when installing the packages, why did you use sudo on
"sudo apt-get install -y openmpi-bin libopenmpi-dev
" and
"sudo apt-get install -y libatlas-base-dev",
my understanding was that when using sudo it installs outside of the environment so are these packages meant to be out of the enviroment? or does it just work differently with apt-get?
I had to read up on this.
I'm pretty sure the sudo command is required for a direct download using apt-get. If I didn't want to use sudo, I could compile the package from source, but that is a few extra commands.
From what I understand, apt-get is a tool to install OS related packages, and virtualenv is a tool to isolate Python environments. So a virtualenv will never be able to isolate packages installed using apt-get.
If you're looking to contain packages downloaded using apt-get, I think "chroot" would be the direction to go. (I haven't used this before but SO told me so.)
ERROR: tensorflow-2.0.0-cp37-none-linux_armv7l.whl is not a supported wheel on this platform.
I followed the instructions but still got this error after several attempts. May I know how to solve it?
I made a 2022 update for this tutorial last week. There are a couple differences so that could be a good place to start ruclips.net/video/QLZWQlg-Pk0/видео.html
why do you require the specific versions for h5py and keras_applications and keras_preprocessing?
Great tutorial! This is very helpful :)
Thank you Montassar!
I have an error of "can't uninstall wrapt" in last command. How should I fix it?
In a virtual environment the --user is not applicable as it doesn't require any specific installation configuration as it's installed within the environment
Hello There,
First of all, thanks for your video!
Secondly, could you pls tell me how to load the virtual environment in Mu or any other code editor?
Thanks
Sandor
Hi I load the virtual environment in the terminal then run code from the terminal
Do you have any recommendations on getting this to install on bullseye?
I am getting a wheel not supported on this platform error.
Some things have changed since I published this video. I recently made an updated video here: ruclips.net/video/QLZWQlg-Pk0/видео.html
Thank you. Hope likes, comment and view are enough to repay your kindness
thanks a lot!!!!!!!!!! almost gave up turns out I was installing the wrong version
Thanks 😎 you are doing gods work my dude had the same emotions when i imported tensorflow on python lol
Happy to help!
can we try this procedure on raspberrypi 3?
Successful on my pi . Thanks
May I ask you,why the system says:[ModuleNotFoundError: there is no module named tensorflow],what is the problem? thank you.
This means your don't have tensorflow installed. Did everything go okay with the installation?
Great job Dear very clear each n everythings. goodluck!! thank you so much
Awesome worked for me tensorflow 2.3.0
Thanks a lot :)
while training my model I'm facing the following error on my terminal!
2020-03-31 19:39:53.793083: W tensorflow/core/framework/cpu_allocator_impl.cc:81] Allocation of 188160000 exceeds 10% of system memory.
Can you please suggest me how to increase my virtual environment memory?
You need to decrease your batch size. This is a memory issue with your device, not environment
@@SamWestby I'd like to know deeper about this problem if you can help me, like if the RAM is responsible or the micro SD card. And if the latter is the issue if I can extend the storage with any USB drive
in 1:23 after I write virtualenv env it says command not found
Thanks Sam! Found this via the picar tutorial.
You're welcome Warren! What is the picar tutorial?
@@SamWestby towardsdatascience.com/deeppicar-part-3-d648b76fc0be you made linked your video in one of the comments. Did you end up making a Pi Car?
That's cool! Here's what I ended up making ruclips.net/video/hJQgWvXT5p0/видео.html
@@SamWestby Rad!
thanks man! you are the best! you saved my life thank you
im kind of new to this, using a pi 3 when I install tensorflow it returns that it is not a supported wheel on this platform any idea what of what I doing wrong?
If you followed my step-by-step tutorial, I'm not sure. Sorry
Hello, I managed to download version 2.0.0, but I need some help in downloading 2.3.0. Can anyone help me with the exact commands
really thank you i have installed from yours, tf version 2.3 on my pi
how ? i have only the 2.0.0 version
An issue I'm facing is that I can't figure out how to program or train a model using TensorFlow in my Pi 4. Can you please help me out? thank you.
Is your issue learning Tensorflow, or is your issue learning Tensorflow for your Pi 4?
@@SamWestby My issue is to learn tensoflow on raspberry Pi 4 as I don't have a windows PC. I am also very eager to learn the concepts of tensoflow also. Also I am keen to know how to make/train a training model. Thanks
I recommend checking out www.tensorflow.org/learn
You don't need a fancy computer to start learning
@@SamWestby I did check and found that I need a full tensoflow version (not lite) to create a model then I can transfer the model to tensoflow lite format. I do require a demo like video where I can create a training model in full tensoflow. So in a sense I am facing touble training models. Apart from that can I update the tensoflow 2.0.0, which you showed in this video to 2.1.0 (the current version) in my Raspberry Pi 4? I have the 4 GB model.
@@SamWestby while training my model I'm facing the following error on my terminal!
2020-03-31 19:39:53.793083: W tensorflow/core/framework/cpu_allocator_impl.cc:81] Allocation of 188160000 exceeds 10% of system memory.
I was following the official tutorial at tensorflow.org as you guided. The problem rose while I was making the def epoch programming. I checked over 10 times but the same prompt was shown to me every time. Can you please guide me through this? Thanks.
Thanks, that worked for me really helped a lot.
Hey Sam, when I use the "virtualenv env" command it says command not found. What should I do?
Try installing virtualenv with pip by running "python3 -m pip install virtualenv"
If it says something along the lines of "installed virtualenv successfully" then "virtualenv env" should work now
@@SamWestby Yeah I got it to work, I had to logout, thank you!
Thank you for sharing!
Thankyou. This works !!
Thanks mate you heleped a lot
Tensorflow is not supported wheel on this platform
Thank u so much bro! Greetings from México!
It works like a charm... thanx mate
You're welcome!
ERROR: Package 'mock' requires a different Python: 3.5.3 not in '>=3.6'
I'm getting the above error. Can you please help me out?
You need a higher version of Python. You need Python 3.6 for that package, but you actually need Python 3.7 for this tutorial
@@SamWestby Okay I did that but now I am getting this error:
ERROR: tensorflow-2.0.0-cp37-none-linux_armv7l.whl is not a supported wheel on this platform.
@@SiddhanthPRao Are you using Python 3.7 and a Raspberry Pi 2-4?
@@SamWestby I'm using the Rpi 3b+ and have upgraded to python 3.7
@@SiddhanthPRao sorry I'm not sure what the error is then :( I'm not an expert at this stuff
Bro, you saved my life rn! Thanks!!
Glad I could help!
Minor correction: for final testing, instead of "tensor.__version__" it should be: "tensorflow.__version__". Otherwise, great tutorial!
Just changed that. Thanks!
Firstly, I used this tutorial in Ubuntu Mate, its difficult to install h5py and grpcio.
Now, install raspOS, python 3.7, follow this tutorial, E A Z Y
Thank You so much dude success :)
great video...... i have founf=d this tutorial after long search
Thanks!
Lifesaver man. Thank you very much.
Glad it helped!
쌤!!! 정말 고마워요!!!
Hey!
Can I install keras, pandas, Opencv, nltk in raspberry pi?
Hi yes you can!
Melissa, yes you can install those with “sudo pip3 install” on Raspbian OS.
Thank you, Sir.
You're welcome!
thanks so much!
you are the best!!!!!
what can I do
sonofagun it worked. thank you!
great videoooo :)!!!
Thanks!!
U should live stream. On twitch
I'll look into it!
THANK IT WOKS FOR ME
Yay!
is this 64 bit OS?
Nope 32 bit raspbian
Thanks man saved me a hell of a time
No working
Спасибо!
Пожалуйста!
bravo bravo
Thank you so much for this.
You are so welcome!
Thank you, work for me..!