Done. Nothing else appeared. Please show next thanks.
root@pc1-desktop:/# python3
Python 3.6.9 (default, Oct 8 2020, 12:12:24)
[GCC 8.4.0] on linux
Type “help”, “copyright”, “credits” or “license” for more information.
import tensorflow
2021-03-04 17:02:36.598499: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.10.2
Next, you can write your own program using tensorflow library. If you do not know how, I suggest you to look at official courses:
https://courses.nvidia.com/courses/course-v1:DLI+S-RX-02+V2/about
Thanks Lissanro. There’s no GUI for TensorFlow?
Yes and no. The course I mentioned explains how to get basic GUI working but of course you will have to write code (or run examples, but you need to write something on your own if you want to learn). There is also tensorboard, very useful tool. And so on. But if you mean GUI which allow you to create your own programs without writing the code, then no, read this if you need more explanation: https://www.quora.com/How-do-I-start-learning-TensorFlow-without-any-coding-experience.
I can copy and paste and edit minor changes all day, but I don’t do coding nor any heavy programming. They should come up with some sort of user friendly GUI. That’s just my 2 cents.
Ok, so I’ve been able to access and run Tensorboard. How do I use it to run specific program, say Metatrader 5? Thanks.
Sry what do you expect to do with the Jetson Nano?
Starting Metatrader 5 in Tensorflow and do “AI” trading for you?
Please make you familiar with tensorflow, pyTorch, docker and python. Programming and linux skills are mandatory.
Yes, that is what I expected. Jetson nano tout itself as AI or machine learning platform right? I’m familiar enough to run and install stuff, but now I actually need to run the AI and machine learning, not just the Python OS.
Python, PyTorch, Tensorflow, Docker, Linux knowledge and the willingness to learn is essential. (This is not a 2 hour -yourleraningplatformhere- thing)
If you willing to learn this is the way you go: jetson-inference/docs at master · dusty-nv/jetson-inference (github.com)
Learn how to grab data from metatrader 5 or how to interact with it.
This is not a Plug and Play Platform which is connecting to a cluster itself and handling some applications.
Thank you for the link. I will see what I need out of it and apply to what is needed for what I do. I don’t need camera nor controller, for example.
You’re welcome.
For learning a camera is great and not really expensive (check the forums for what is supported, a webcam works aswell).
For achieving a “cluster”-compute you may check and learn “docker swarm” or better “kubernetes”.
why there is an error when I install the lib ‘h5py’?
What is the error, what version of h5py you are trying to install?
Thank you Xerox for letting me know about Kubernetes! It’s like brainmapping, spreadsheet, and macros all rolled into one except for containers! And it upgrades these containers automatically! I’m novice with AI stuff but certainly not brainmapping, spreadsheet or macro. I will fully utilize the Kubernetes. I’m currently enrolled in MS Azure Devop but I will figure out others. Thanks again!
Just run pip install h5py.
I first try to run it on Jetpack4.3, and there is an error.
Then I run it on Jetpack4.4, the error still exists.
I tested out tensorflow on Nano (4GB) today with https://github.com/HENDRIX-ZT2/Deep-Video-Deinterlacing which results in 1.5FPS while GPU is at 100% permanently for 576i input. Is this what is to be expected or am I missing anything? (I know, the repo uses ffmpeg but I presume that’s not the reason for the low FPS)
I have no idea what error you are talking about, so I can only try to guess. For me, when I was using JetPack 4.4 and tried to sudo -H pip3 install h5py==2.9.0, it failed, you can look at the log to see if the error is the same in your case. But running sudo -H pip3 install h5py==2.10.0 finished successfully. Installing h5py this way also worked with JetPack 4.5.
Hello I am a fresh to jetson nano also for docker. I try some ways from Internet but failed. so I have to ask for your help
my question is after my installation of the TensorFlow how can I save the docker environment to use next time.
because I find every time I restart the docker the tensorflow is missing. thanks for your help !
I think I need more study of docker .
Hello there, I tried to update the tensorflow version to 2.1.0 on my JetBot with JetPack Version 4.3 by following the official NVIDIA guide.
However it fails and gives me the following error:
jetbot@jetson-4-3:~$ sudo pip3 install --pre --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v43 tensorflow==2.1.0+nv20.2
Looking in indexes: https://pypi.org/simple, https://developer.download.nvidia.com/compute/redist/jp/v43
Collecting tensorflow==2.1.0+nv20.2
Using cached https://developer.download.nvidia.com/compute/redist/jp/v43/tensorflow/tensorflow-2.1.0%2Bnv20.2-cp36-cp36m-linux_aarch64.whl (236.9 MB)
WARNING: Discarding https://developer.download.nvidia.com/compute/redist/jp/v43/tensorflow/tensorflow-2.1.0%2Bnv20.2-cp36-cp36m-linux_aarch64.whl (from https://developer.download.nvidia.com/compute/redist/jp/v43/tensorflow/). Requested tensorflow==2.1.0+nv20.2 from https://developer.download.nvidia.com/compute/redist/jp/v43/tensorflow/tensorflow-2.1.0%2Bnv20.2-cp36-cp36m-linux_aarch64.whl has inconsistent version: filename has '2.1.0+nv20.2', but metadata has '2.1.0+nv20.2.tf2'
ERROR: Could not find a version that satisfies the requirement tensorflow==2.1.0+nv20.2 (from versions: 1.15.2+nv20.2, 1.15.2+nv20.3, 2.1.0+nv20.2, 2.1.0+nv20.3)
ERROR: No matching distribution found for tensorflow==2.1.0+nv20.2
I don’t really know how to fix this, any help is appreciated …
Thanks