Thank you so much. I have been struggling with tensorflow-gpu setup errors for days. Your videos have really helped me. I finally got it to work. I am also a huge fan of your playlist on Deep Neural Networks. It has been very helpful to me. Thanks again.
I screwed up my Tensorflow environment this weekend. Followed Jeff’s instructions and was going again in under an hour!!!! Last time I think it took 6! These are awesome
Thank you professor. I've been trying to do this for weeks, but I failed every time. But when I followed your instructions today, it worked like magic! I also have a friend who is having trouble installing, and I sent him the link :) Thank you so much!!!
tried it in anaconda 3.8 - no gpu. Went for Minicoda, the exact 3.7 one in the video, followed everything. GPU is available :) Maximum joy. Thanks Jeff.
Thank you Jeff! I have tried so many other tutorials on this issue, but none of them worked... Your Tutorial is realy simple and straigt to the point with some useful information!
"conda install -c anaconda tensorflow-gpu" Doesn't work for me. It shows "PackagesNotFoundError: The following packages are not available from current channels:", Please help me.
On TenforFlow's install GPU support page it says you no long need to use the -gpu, so instead try "conda install -c anaconda tensorflow" as this should work for both GPU and CPU
hi, I'm using nvidia geforce gtx 1050ti, I also already installed the driver, cuda toolkit, and cudnn. I've tried to install TF via anaconda, but it always cannot read my gpu. I've also test it using your code, Tensor Flow Version: 2.3.0 Keras Version: 2.4.0 Python 3.8.10 (default, May 19 2021, 13:12:57) [MSC v.1916 64 bit (AMD64)] GPU is NOT AVAILABLE could you help me with that? Thank you
It should work. I would suggest trying to get to the exact error it is reporting and google that. I do not have experience with this specific GPU, though.
I have always hated this environment setting and almost never did it right way. Thus ending up installing TF-CPU, this time around it was done the right way. Thank you.
When jupyter open, it doesn't look like this 15:48. And when i click to zip, it errors: Error! C:\Users\admin\Downloads\Miniconda3-py39_4.9.2-Windows-x86_64.exe is not UTF-8 encoded Saving disabled. See Console for more details.
thank you professor, this is super helpful edit: as a note, the gpu didn't work for me with python 3.8 but it worked when i changed it to 3.7.7 [windows]
hey sir, every time i follow steps this error which is about missing DLL files ``` >>> import sys >>> print(f"Python {sys.version}") Python 3.8.5 (default, Sep 3 2020, 21:29:08) [MSC v.1916 64 bit (AMD64)] >>> gpu = len(tf.config.list_physical_devices('GPU'))>0 2020-10-01 16:28:09.086099: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library nvcuda.dll 2020-10-01 16:28:09.115435: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:01:00.0 name: GeForce GTX 1660 Ti computeCapability: 7.5 coreClock: 1.59GHz coreCount: 24 deviceMemorySize: 6.00GiB deviceMemoryBandwidth: 268.26GiB/s 2020-10-01 16:28:09.115645: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudart64_101.dll 2020-10-01 16:28:09.118208: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cublas64_10.dll'; dlerror: cublas64_10.dll not found 2020-10-01 16:28:09.119406: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cufft64_10.dll'; dlerror: cufft64_10.dll not found 2020-10-01 16:28:09.120489: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'curand64_10.dll'; dlerror: curand64_10.dll not found 2020-10-01 16:28:09.121608: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cusolver64_10.dll'; dlerror: cusolver64_10.dll not found 2020-10-01 16:28:09.140394: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cusparse64_10.dll 2020-10-01 16:28:09.190095: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudnn64_7.dll 2020-10-01 16:28:09.190337: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1753] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform. Skipping registering GPU devices... >>> print("GPU is", "available" if gpu else "NOT AVAILABLE") GPU is NOT AVAILABLE ```
gud morning Sir jeff I successfully installed tensorflow with gpu but when i install addtnl librriaries here is the error Warning: you have pip-installed dependencies in your environment file, but you do not list pip itself as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging you. pls help tnx
I had the same problem and solved it like explained here stackoverflow.com/questions/67081266/conda-yml-file-prompts-please-add-an-explicit-pip-dependency . Hope it helps you.
Hello, I'm getting a "The procedure entry point could not be located in the dynamic link library" Error when trying to register my environment. Something related to this path C:\Users\Me\miniconda3\Library\bin\pythoncom38.dll
Wow thank you so much.. it all actually worked! I started the process fully expecting to run to Colabs but it all went fine. Thankfully the gpu all worked fine too without errors and the need for drivers. I'm running a GeForce GTX 1660 Ti.
Thank you for the tutorial. I have an issue at the: "conda env update --file tools.ym" step. The cmd floods with "Warning: you have pip-installed dependencies in your environment file, but you do not list pip itself as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging you." - in an endless loop.
conda install -y jup says "conda is not recognized as an internal or external command" .. when done inside of the anaconda prompt, it says it is collecting package metadata, then says "packagesnotfounderror: the following packages are not available from the current channels" - jupiter" ... tried various different spellings, jup, jypiter, jupiter, all same error
Hello Jeff, I installed tensorflow-gpu version and the installation ran well and it looked like everything was working but I got the message GPU is NOT AVAILABLE when I ran the jupyter notebook, could you help me?
Thanks for the video, after days of trying I finally managed to get the gpu recognised as a device in spyder. I wonder if you can help further. I'm getting the below error in jupyter when I try fit a model. ~\anaconda3\envs\tensorflow\lib\site-packages\tensorflow_core\python\eager\execute.py in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name) 65 else: 66 message = e.message ---> 67 six.raise_from(core._status_to_exception(e.code, message), None) 68 except TypeError as e: 69 keras_symbolic_tensors = [ ~\AppData\Roaming\Python\Python37\site-packages\six.py in raise_from(value, from_value) UnknownError: [_Derived_] Fail to find the dnn implementation. [[{{node CudnnRNN}}]] [[sequential/gru/StatefulPartitionedCall]] [[Reshape_9/_26]] [Op:__inference_distributed_function_2575] Function call stack: distributed_function -> distributed_function -> distributed_function
Did everything as you did in the Video, but this comes up: Tensor Flow Version: 2.3.0 Keras Version: 2.4.0 Pythin 3.8.5 Pandas 1.2.2 Scikit-Learn 0.23.2 GPU is NOT AVAIBLE Im on a GTX 1080. Any ideas?
Спасибо большое за видео! Очень подробный гайд, к сожалению не смог установить gpu версию Tensorflow, но думаю, что на первое время хватит и обычной cpu версии. Удачи!
Install Additional Libraries for ML; (tensorflow) C:\Users\LAPTOP>conda env update --file tools.yml EnvironmentFileNotFound: 'C:\Users\LAPTOP\tools.yml' file not found Any Idea on how to proceed here, i am stuck with installing additional libraries :( (Python 3.9 all ready installed)
Tensor Flow Version: 2.4.0 Keras Version: 2.4.0 Python 3.8.5 (default, Sep 4 2020, 02:22:02) [Clang 10.0.0 ] Pandas 1.1.5 Scikit-Learn 0.23.2 GPU is NOT AVAILABLE I have an NVIDIA GeForce GTX 1060 6 GB
Thanks - it worked with python=3.7, but does not seem to work with python=3.8 environment - no GPU is showing up, even though TF v. 2.4 should work fine with 3.8 ...
Ok, I did this: python -m ipykernel install --user --name tensorflow --display-name "Python 3.7 (tensorflow)" some times with different names and now I have a gazillion kernels... how do I remove the unwanted ones?
Hey, anaconda latest version is available with python 3.8 but TensorFlow 2.1 does not support this python version so how can we install anaconda with python 3.7 only and please don't tell me to create a venv.
I went through the entire process's best explanation. but the auto-completion code is not working in the new environment I tried nb_extensions also. can you help?
Thanks a lot for the tutorial. I used anaconda instead and noticed that if i use the command prompt or powershell i get a lot of issues (conflicts). Using the anaconda prompt instead resolved everything. Just thought sharing this, and again many thanks to Jeff for the tutorial, well presented.
Hi Jeff. I see that you do a fresh Win 10 environment, install your stuff, play around, and then toss it when you don't need it or things change. For those of us that can't do that would your installation instructions work in a docker Win 10 environment as well? I'm not sure how the CUDA drivers install would work with Docker. Thanks for sharing your knowledge and making the effort to keep your content as up to date as possible :-)
I happened to have just built a machine. However, I usually test on a clean virtual box just to be sure I am not skipping something that was already setup. I do have a Docker image, GPU is a bit more tricky with Docker. I've done Linux hosts with NVIDIA-Docker, not entirely sure that works on Windows. Probably would with WSL. Here is my docker image. hub.docker.com/r/heatonresearch/jupyter-python-r
I could not downloaad "t81_558_deep_learning" folder from your given link, It doesn't show file size and continuously donwloading. I used IDM for downloading. please help me.
Installation went well, except GPU support (Geforce GTX 1050) - gave me error at the last (validation) step. What helped me is just installing latest supported NVIDIA driver for my card (after updating driver I just re-ran the validation code and GPU was found).
The error I am getting now at the command where we use python to install the ipykernal is this: ImportError: cannot import name 'generator_to_async_generator' from 'prompt_toolkit.eventloop' (C:\Users\Lowz\Anaconda3\envs\tensorflow\lib\site-packages\prompt_toolkit\eventloop\__init__.py) What is happening here?? EDIT: It seems there is an error right now with the ipython and prompt-toolkit versions being incompatible. ipython 7.17.0 requires prompt-toolkit to be less than 3.0.01 but greater than 2.0.0, but jupyter-console requires prompt-toolkit version less than 2.0.0 but greater than 1.0.0. So...how can this be possible to work around? EDIT2: Okay for those of you using python 3.7.7, if you just change the "Python 3.7" flag in the command to "Python 3.7.7" it should work -it did for me after uninstalling prompt-toolkit and just reinstalling the latest version which was 3.0.6 at this time.
I have an error at the point were you install the additional libraries for ML. The line used in command prompt is "conda env update --file tools.yml" All the stepts up until this point worked fine but once i put this in the command prompt in said that "EnvironmentFileNotFound: 'C:\Users\User\tools.yml' file not found" If anyone has a solution to this please post it in the reply comment to this comment. Thanks.
Dear Professor, I have installed tensorflow in anaconda (not miniconda). I have done my image segmentation using python 3 but when I run tensorflow for deep learning (CNN-LSTM or convLSTM), a pop up comes mentioning that the kernel has died and it will activate automatically. However, it does not activate. I tried my best to fix this but was unsuccessful. I would be really grateful for your kind advice as I have to run deep learning models for my PhD
Heya, I seem to have a little problem. When I go to install the tools.yml the cmd fills with "Conda may not use the correct pip to install your package, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging.". It has been doing this for a while is this normal?
I just skipped this step and in the end when I run the scan I received some import errors. To solve it, go into the anaconda navigator -> environments -> tensorflow and looked for "pandas" and "scikit-learn" in not installed. Run the verification again and all be fine. Remember to use python 3.7.10 (latest python 3.7 version) to use GPU in tensorflow. Edit1: To donwgrade python run: conda install python=3.7 in your Prompt or anaconda Prompt
I tried this on the 18th of December 2020. By now there is a new version of python available 2.8. I changed the version of python in the setup of the environment from 2.7 to 2.8. Everything installed without any error but the final check told me that the GPU is not available. So I repeated the process with the 2.7 version and everything went fine. GPU is now available. On a side note: I like how easy this was compared to when I did this with the full driver setup.
This is exactly the problem I had. It told me "GPU NOT AVAILABLE" when I finished. As of Dec 2020, TF is not compatible with Python 3.8 or 3.9. You must use 3.7 to avoid this problem.
13:44 if didn't work, the files you downloaded transfer it to the specific director...mine is "this PC > LOCAL C > WINDOWS > SYSTEM 32", and then it work...
"AttributeError: module 'tensorflow_core.python.keras.api._v2.keras.activations' has no attribute 'swish' " i got this error when i tried to run my NLP code that requires CUDA
Hi Jeff, I got an error while trying to install the tensorflow-gpu version as you did in the video above. See below for error: InvalidArchiveError('Error with archive C:\\Users\\Ivan Aerlic\\.conda\\pkgs\\cudatoolkit-10.1.243-h74a9793_0.tar.bz2. You probably need to delete and re-download or re-create this file. Message from libarchive was:
bzip decompression failed'). Please let me know what you think when you have a spare moment. Thank you.
After using (conda env update --file tools.yml) I get a infinite error: Warning: you have pip-installed dependencies in your enviornment file, but you do not list pip instelf as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. please add an explicit pip dependency. I'm adding one for you, but still nagging you.
@@HeatonResearch a video about linux based systems with cuda would be much appreciated (but no pressure i might be able to figure it out on my own). the problem im having is i dont have libcublas.so.10.0 packed with the rest of my cuda library. currently i believe that this problem is due to the fact that i am using python3.8, which only supports tensorflow version 2.2.0 or 2.3.0 (not 2.0.0), both of which support cuda-10.1 or higher. since the release of cuda-10.1, libcublas has not been prepackaged with the rest of the cuda library. correct me if i am wrong here, but should i be using python3.7 so that i can use tensorflow 2.0.0 which supports cuda-10.0 (this is the latest version with libcublas prepackaged)?
I get this: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above. [[node model/conv1d/conv1d (defined at :3) ]] [Op:__inference_distributed_function_1394] Function call stack: distributed_function when i try to train one epoch with a cnn =(
Sir, please can u suggest me the current CUDA and cuDNN versions for my GPU (GTX 1060 6gb) ? I have tried a lot.. in the end my gpu is not detected by TF ..
Hi Jeff, I am using TensorFlow version 1.0.0 with python 3.5 to train my RNN model on Anaconda spyder. But now since I started the training I found out that it maxes out my CPU. I have an NVIDIA graphics card GTX 1050 and want to use it with minimal changes in my code. How can I do this.
Hello jeff, I'm starting in this world of neuronal networks, I have a question about of this topic: Do you know if is possible to configurate like as gpu an integrate gpu how is AMD with Radeon vega 11? I configurated the enviroment wirh cpu but i want to know if is possible configurate that GPU. Thanks!
"conda install -c anaconda tensorflow-gpu" Doesn't work for windows. do you help me for making a correction? It also error in "PackagesNotFoundError: The following packages are not available from current channels:". Please help me.
I have been having a hard time with conda installation on my PC. It don't seem to show the navigation menu anywhere and when I click on the Anaconda cmd, it opens and closes almost immediately. I've tried different suggestions but it hasn't been better, will love to hear from you also
Thanks for the vid! I had a problem where when I tried to run the t81_558_class_01_1_overview on jupyter notebook, this error occurs: AttributeError: module 'tensorflow' has no attribute 'compat'. Id appreciate any help!
While trying to install miniconda, I see Python 3.8 and python 2.7. In this video, it shows Python 3.7. Its a problem because I read at places that tensorflow doesn't support Python 3.8.How should I proceed?
hi sir, I have a question. when startup deploys their models on cloud, I don't think they would be running instance 24*365 it would be very costly. can you please explain how they do it?
The cloud can get costly. My only suggestion is to use automation and use for only as much time as is needed. Also do not use more hardware than you need. Multiple GPUs won't help some models.
I have been following this and there was no DLL error. Tensorflow recognize my RTX 2070 GPU and used all of its memory but during training I open my task manager window and see that CUDA is not working
great video sir, I just want to know. if I installed python ver 3.9 and im using the latest miniconda3 that use python 3.8. will it affect the miniconda if I tick the box of "add miniconda3 to my PATH environment variable".??
at 8:52 , you use python= 3.7 because tensorflow required to use that version or because based on miniconda python version u downloaded?. mine use python 3.8 for the miniconda. should I write conda create --name tensorflow python=3.8??
Thank you for your videos, looking for your help guys, I installed Python 3.8 on my Windows when I got to install tensorflow with the command " conda install -c anaconda tensorflow" an error says its incompatible, your advice please
Thanks so much for the useful video I followed everything and it appeared that tensorflow uses GPU but every time I try to continue a message appears in jupyter NB : Kernel appears to have died it will restart automatically , Also this error appears : Attempting to fetch value instead of handling error Internal: failed to get device attribute 13 for device 0: CUDA_ERROR_UNKNOWN: unknown error I don't know why .. so could you help me solving such problem ? thank you so much
Thank you so much. I have been struggling with tensorflow-gpu setup errors for days. Your videos have really helped me. I finally got it to work.
I am also a huge fan of your playlist on Deep Neural Networks. It has been very helpful to me. Thanks again.
Glad it helped! Thanks for your comment.
is it possible for us to run pythoncodes using gpu but without gpu, am stuck can you help
I screwed up my Tensorflow environment this weekend. Followed Jeff’s instructions and was going again in under an hour!!!!
Last time I think it took 6!
These are awesome
Thank You for that video, I've tried a lot of guides, but only your helped me to configure tensorflow with GPU with anaconda and jupyter!
You, sir, are the best explaining how to do this. For the first time, I did successfully installed Tensorflow for GPU. I am subscribed now
Thank you professor. I've been trying to do this for weeks, but I failed every time. But when I followed your instructions today, it worked like magic! I also have a friend who is having trouble installing, and I sent him the link :) Thank you so much!!!
Glad it was helpful!
which python version you choose? 3.8 or 3.7?
@@PortesNew I chose 3.7.. I don't think 3.6 works
tried it in anaconda 3.8 - no gpu. Went for Minicoda, the exact 3.7 one in the video, followed everything. GPU is available :)
Maximum joy. Thanks Jeff.
Concrete and precise ! Loved the way you walked through the entire process.
sir i have phython version 3.9 and I am getting an error the version is not supportable with tensorflow
Glad to see that prof is still next to the backbone of the internet lol
Thanks!
@@HeatonResearch Can you help me to debug? I post the error I got in a comment
Thank you Jeff! I have tried so many other tutorials on this issue, but none of them worked... Your Tutorial is realy simple and straigt to the point with some useful information!
bruh this is actual hell, I hate doing this stuff. Thanks for making it easier
Thank you sir for making such an informative video. I finally now have tensorflow on my system after 2 days of struggle
"conda install -c anaconda tensorflow-gpu" Doesn't work for me. It shows "PackagesNotFoundError: The following packages are not available from current channels:", Please help me.
Check your Python version maybe
On TenforFlow's install GPU support page it says you no long need to use the -gpu, so instead try "conda install -c anaconda tensorflow" as this should work for both GPU and CPU
Thank you for posting this, recently shifted to windows os and your yt channel has been a life saver
conda install nb_conda wont work for python 3.9 and higher
hi, I'm using nvidia geforce gtx 1050ti, I also already installed the driver, cuda toolkit, and cudnn. I've tried to install TF via anaconda, but it always cannot read my gpu. I've also test it using your code,
Tensor Flow Version: 2.3.0
Keras Version: 2.4.0
Python 3.8.10 (default, May 19 2021, 13:12:57) [MSC v.1916 64 bit (AMD64)]
GPU is NOT AVAILABLE
could you help me with that? Thank you
It should work. I would suggest trying to get to the exact error it is reporting and google that. I do not have experience with this specific GPU, though.
I have always hated this environment setting and almost never did it right way. Thus ending up installing TF-CPU, this time around it was done the right way.
Thank you.
When jupyter open, it doesn't look like this 15:48. And when i click to zip, it errors:
Error! C:\Users\admin\Downloads\Miniconda3-py39_4.9.2-Windows-x86_64.exe is not UTF-8 encoded
Saving disabled.
See Console for more details.
if i run code script that u share on the github page , "GPU is NOT AVAILABLE" how to solve it sir ?
I have the same issue, I have an NVIDIA GeForce GTX 1060. What about you?
Thank you so much! I've been looking for this for ages!
Glad I could help!
thank you professor, this is super helpful
edit: as a note, the gpu didn't work for me with python 3.8 but it worked when i changed it to 3.7.7 [windows]
hey sir, every time i follow steps this error which is about missing DLL files
``` >>> import sys
>>> print(f"Python {sys.version}")
Python 3.8.5 (default, Sep 3 2020, 21:29:08) [MSC v.1916 64 bit (AMD64)]
>>> gpu = len(tf.config.list_physical_devices('GPU'))>0
2020-10-01 16:28:09.086099: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library nvcuda.dll
2020-10-01 16:28:09.115435: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce GTX 1660 Ti computeCapability: 7.5
coreClock: 1.59GHz coreCount: 24 deviceMemorySize: 6.00GiB deviceMemoryBandwidth: 268.26GiB/s
2020-10-01 16:28:09.115645: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudart64_101.dll
2020-10-01 16:28:09.118208: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cublas64_10.dll'; dlerror: cublas64_10.dll not found
2020-10-01 16:28:09.119406: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cufft64_10.dll'; dlerror: cufft64_10.dll not found
2020-10-01 16:28:09.120489: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'curand64_10.dll'; dlerror: curand64_10.dll not found
2020-10-01 16:28:09.121608: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'cusolver64_10.dll'; dlerror: cusolver64_10.dll not found
2020-10-01 16:28:09.140394: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cusparse64_10.dll
2020-10-01 16:28:09.190095: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudnn64_7.dll
2020-10-01 16:28:09.190337: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1753] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.
Skipping registering GPU devices...
>>> print("GPU is", "available" if gpu else "NOT AVAILABLE")
GPU is NOT AVAILABLE
```
FINALLY, I GOT IT RUNNING AFTER LOTS OF TUTORIALS I TRIED! THANKS
gud morning Sir jeff I successfully installed tensorflow with gpu but when i install addtnl librriaries here is the error Warning: you have pip-installed dependencies in your environment file, but you do not list pip itself as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging you. pls help tnx
I had the same problem and solved it like explained here stackoverflow.com/questions/67081266/conda-yml-file-prompts-please-add-an-explicit-pip-dependency . Hope it helps you.
I tried following this tutorial but jupyter notebook just opens a blank tab when I use it
Very helpful sir ...can you help me out for my phd related work on retinal segmentation
That sounds very interesting; however, my bandwidth for side projects is not good.
Hello,
I'm getting a "The procedure entry point could not be located in the dynamic link library" Error when trying to register my environment. Something related to this path C:\Users\Me\miniconda3\Library\bin\pythoncom38.dll
sir i could not install python 3.7 because it is not available on the website now
Wow thank you so much.. it all actually worked! I started the process fully expecting to run to Colabs but it all went fine. Thankfully the gpu all worked fine too without errors and the need for drivers. I'm running a GeForce GTX 1660 Ti.
@Aashik Sharif B I didn't need to install any drivers. 😁
I am getting "GPU is NOT AVAILABLE" yet I have everything in place for my RTX 2070 and I did use "conda install -c anaconda tensorflow-gpu"
Thank you very much, it really helped me to install tensorflow for gpu quickly, after several hours trying other tutorials out there!
Thank you for the tutorial. I have an issue at the: "conda env update --file tools.ym" step. The cmd floods with "Warning: you have pip-installed dependencies in your environment file, but you do not list pip itself as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging you." - in an endless loop.
I have the same problem with yours
Jeff I cannot thank you enough for this video! You've helped so many people
conda install -y jup says "conda is not recognized as an internal or external command" .. when done inside of the anaconda prompt, it says it is collecting package metadata, then says "packagesnotfounderror: the following packages are not available from the current channels" - jupiter" ... tried various different spellings, jup, jypiter, jupiter, all same error
Thanks for the video, is helping getting the job done!
just a note:
pip is no longer included in conda installer, so you should modify the tools.yml adding pip:
# conda create --name tensorflow-cpu python=3.8
# conda install jupyter
# conda install nb_conda
# conda install -c anaconda tensorflow
# conda env update --file tools.yml
dependencies:
- pip
- jupyter
- scikit-learn
- scipy
- pandas
- pandas-datareader
- matplotlib
- pillow
- tqdm
- requests
- h5py
- pyyaml
- flask
- boto3
- pip:
- bayesian-optimization
- gym
- kaggle
Hello Jeff, I installed tensorflow-gpu version and the installation ran well and it looked like everything was working but I got the message GPU is NOT AVAILABLE when I ran the jupyter notebook, could you help me?
Thanks for the video, after days of trying I finally managed to get the gpu recognised as a device in spyder.
I wonder if you can help further. I'm getting the below error in jupyter when I try fit a model.
~\anaconda3\envs\tensorflow\lib\site-packages\tensorflow_core\python\eager\execute.py in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
65 else:
66 message = e.message
---> 67 six.raise_from(core._status_to_exception(e.code, message), None)
68 except TypeError as e:
69 keras_symbolic_tensors = [
~\AppData\Roaming\Python\Python37\site-packages\six.py in raise_from(value, from_value)
UnknownError: [_Derived_] Fail to find the dnn implementation.
[[{{node CudnnRNN}}]]
[[sequential/gru/StatefulPartitionedCall]]
[[Reshape_9/_26]] [Op:__inference_distributed_function_2575]
Function call stack:
distributed_function -> distributed_function -> distributed_function
solved with this:
import tensorflow as tf
physical_devices = tf.config.list_physical_devices('GPU')
tf.config.experimental.set_memory_growth(physical_devices[0], enable=True)
Did everything as you did in the Video, but this comes up:
Tensor Flow Version: 2.3.0
Keras Version: 2.4.0
Pythin 3.8.5
Pandas 1.2.2
Scikit-Learn 0.23.2
GPU is NOT AVAIBLE
Im on a GTX 1080. Any ideas?
I got the same issue, did you fix this?
Спасибо большое за видео! Очень подробный гайд, к сожалению не смог установить gpu версию Tensorflow, но думаю, что на первое время хватит и обычной cpu версии. Удачи!
Thank you very much sir after 6 hours of trial and error this worked perfectly for me :D
Glad it helped.
also help it cmd prompt will say conda is not recognized as a command
wait nvm i used the miniconda prompt
@@etrieusvet2070 Yeah that is a common problem. Glad you found it!
Thank you very much for this tutorial. I was able to install tensorflow without any issues. This would be one of the most helpful videos.
How to activate conda environment from python script?
Install Additional Libraries for ML;
(tensorflow) C:\Users\LAPTOP>conda env update --file tools.yml
EnvironmentFileNotFound: 'C:\Users\LAPTOP\tools.yml' file not found
Any Idea on how to proceed here, i am stuck with installing additional libraries :( (Python 3.9 all ready installed)
[WinError 87] The parameter is incorrect at cmd, anyone know what my problem?
i have installed tensorflow gpu same but i am getting error no module named keras...but is it necessary to download again keras directly?
hi Professor. Where can I get the download link for Window Miniconda Python3.7 ? they got only the 3.8 version on the website
I have question, what if i install "conda install -c anaconda tensorflow", can i change to "conda install -c anaconda tensorflow-gpu"? ?
whats the speed of colab and its limitation ?
Thx Prof. for your directions. It helps a lot. BTW, it seems that Python 3.8 + CUDA 10.1 + cuDNN 7.6 do not work with Tensorflow-gpu 2.3.0...
yep I don't think it works, im getting an error that says not module named 'flatbuffers' found
Tensor Flow Version: 2.4.0
Keras Version: 2.4.0
Python 3.8.5 (default, Sep 4 2020, 02:22:02)
[Clang 10.0.0 ]
Pandas 1.1.5
Scikit-Learn 0.23.2
GPU is NOT AVAILABLE
I have an NVIDIA GeForce GTX 1060 6 GB
Thanks - it worked with python=3.7, but does not seem to work with python=3.8 environment - no GPU is showing up, even though TF v. 2.4 should work fine with 3.8 ...
Ok, I did this: python -m ipykernel install --user --name tensorflow --display-name "Python 3.7 (tensorflow)" some times with different names and now I have a gazillion kernels... how do I remove the unwanted ones?
I found it :)
Hey, anaconda latest version is available with python 3.8 but TensorFlow 2.1 does not support this python version so how can we install anaconda with python 3.7 only and please don't tell me to create a venv.
hey bro did you get the answer for that problem i am facing the same issue
I went through the entire process's best explanation. but the auto-completion code is not working in the new environment I tried nb_extensions also. can you help?
Thanks a lot for the tutorial. I used anaconda instead and noticed that if i use the command prompt or powershell i get a lot of issues (conflicts). Using the anaconda prompt instead resolved everything. Just thought sharing this, and again many thanks to Jeff for the tutorial, well presented.
hey I faced the same issue and solved it with your suggestion! Many thanks!
Please sir, how will know whether I am using gpu or cpu
Hi Jeff. I see that you do a fresh Win 10 environment, install your stuff, play around, and then toss it when you don't need it or things change. For those of us that can't do that would your installation instructions work in a docker Win 10 environment as well? I'm not sure how the CUDA drivers install would work with Docker. Thanks for sharing your knowledge and making the effort to keep your content as up to date as possible :-)
I happened to have just built a machine. However, I usually test on a clean virtual box just to be sure I am not skipping something that was already setup. I do have a Docker image, GPU is a bit more tricky with Docker. I've done Linux hosts with NVIDIA-Docker, not entirely sure that works on Windows. Probably would with WSL. Here is my docker image. hub.docker.com/r/heatonresearch/jupyter-python-r
@@HeatonResearch Excellent, thanks for replying.
I could not downloaad "t81_558_deep_learning" folder from your given link, It doesn't show file size and continuously donwloading. I used IDM for downloading. please help me.
thanks Jeff, so helpful, no issues at all amazing.
Great to hear!
Installation went well, except GPU support (Geforce GTX 1050) - gave me error at the last (validation) step. What helped me is just installing latest supported NVIDIA driver for my card (after updating driver I just re-ran the validation code and GPU was found).
Good suggestion.
is it possible for us to run python codes using gpu but without virtual environment am stuck can you help
Really good video, it was very in-depth and nice. Thanks for sharing.
The error I am getting now at the command where we use python to install the ipykernal is this:
ImportError: cannot import name 'generator_to_async_generator' from 'prompt_toolkit.eventloop' (C:\Users\Lowz\Anaconda3\envs\tensorflow\lib\site-packages\prompt_toolkit\eventloop\__init__.py)
What is happening here??
EDIT: It seems there is an error right now with the ipython and prompt-toolkit versions being incompatible. ipython 7.17.0 requires prompt-toolkit to be less than 3.0.01 but greater than 2.0.0, but jupyter-console requires prompt-toolkit version less than 2.0.0 but greater than 1.0.0. So...how can this be possible to work around?
EDIT2: Okay for those of you using python 3.7.7, if you just change the "Python 3.7" flag in the command to "Python 3.7.7" it should work -it did for me after uninstalling prompt-toolkit and just reinstalling the latest version which was 3.0.6 at this time.
I have an error at the point were you install the additional libraries for ML. The line used in command prompt is "conda env update --file tools.yml" All the stepts up until this point worked fine but once i put this in the command prompt in said that "EnvironmentFileNotFound: 'C:\Users\User\tools.yml' file not found" If anyone has a solution to this please post it in the reply comment to this comment. Thanks.
I copied the file 'tools.yml' to my active directory, and it worked. Make sure you have write access for that directory.
@@stefanrakuff1425 Thanks for the help have good day.
Dear Professor, I have installed tensorflow in anaconda (not miniconda). I have done my image segmentation using python 3 but when I run tensorflow for deep learning (CNN-LSTM or convLSTM), a pop up comes mentioning that the kernel has died and it will activate automatically. However, it does not activate. I tried my best to fix this but was unsuccessful. I would be really grateful for your kind advice as I have to run deep learning models for my PhD
Heya, I seem to have a little problem. When I go to install the tools.yml the cmd fills with "Conda may not use the correct pip to install your package, and they may end up in the wrong place. Please add an explicit pip dependency. I'm adding one for you, but still nagging.". It has been doing this for a while is this normal?
I just skipped this step and in the end when I run the scan I received some import errors. To solve it, go into the anaconda navigator -> environments -> tensorflow and looked for "pandas" and "scikit-learn" in not installed. Run the verification again and all be fine. Remember to use python 3.7.10 (latest python 3.7 version) to use GPU in tensorflow.
Edit1: To donwgrade python run: conda install python=3.7 in your Prompt or anaconda Prompt
Thank you so much Prof. Heaton
I tried this on the 18th of December 2020. By now there is a new version of python available 2.8. I changed the version of python in the setup of the environment from 2.7 to 2.8. Everything installed without any error but the final check told me that the GPU is not available. So I repeated the process with the 2.7 version and everything went fine. GPU is now available. On a side note: I like how easy this was compared to when I did this with the full driver setup.
This is exactly the problem I had. It told me "GPU NOT AVAILABLE" when I finished. As of Dec 2020, TF is not compatible with Python 3.8 or 3.9. You must use 3.7 to avoid this problem.
13:44 if didn't work, the files you downloaded transfer it to the specific director...mine is "this PC > LOCAL C > WINDOWS > SYSTEM 32", and then it work...
your videos are very helpful always
"AttributeError: module 'tensorflow_core.python.keras.api._v2.keras.activations' has no attribute 'swish' " i got this error when i tried to run my NLP code that requires CUDA
Hi Jeff, I got an error while trying to install the tensorflow-gpu version as you did in the video above. See below for error:
InvalidArchiveError('Error with archive C:\\Users\\Ivan Aerlic\\.conda\\pkgs\\cudatoolkit-10.1.243-h74a9793_0.tar.bz2. You probably need to delete and re-download or re-create this file. Message from libarchive was:
bzip decompression failed'). Please let me know what you think when you have a spare moment. Thank you.
After using (conda env update --file tools.yml) I get a infinite error:
Warning: you have pip-installed dependencies in your enviornment file, but you do not list pip instelf as one of your conda dependencies. Conda may not use the correct pip to install your packages, and they may end up in the wrong place. please add an explicit pip dependency. I'm adding one for you, but still nagging you.
Hi i have a problem when using keras with this. the error says keras requires tensorflow 2.2 and above . Could u help me?
I have the same problem. Did you get any solution?
This is so helpful professor! Thank you for this and your amazing course! :D
Thank you, glad it was helpful!
hello! did you ever end up making the video on how to install the cuda drivers for a linux dual boot on windows 10? :)
This man needs to make a rocket. Said video would be greatly appreciated
@@zaneflory Right! I will get on that. I've actually been wanting to do that since about age 10. :-)
Not yet, dual boot linux and configuring local linux for AI is coming, though.
@@HeatonResearch Thanks! Me and Bennett here are working together on a project and TensorFlow is giving him many many issues haha
@@HeatonResearch a video about linux based systems with cuda would be much appreciated (but no pressure i might be able to figure it out on my own).
the problem im having is i dont have libcublas.so.10.0 packed with the rest of my cuda library. currently i believe that this problem is due to the fact that i am using python3.8, which only supports tensorflow version 2.2.0 or 2.3.0 (not 2.0.0), both of which support cuda-10.1 or higher. since the release of cuda-10.1, libcublas has not been prepackaged with the rest of the cuda library.
correct me if i am wrong here, but should i be using python3.7 so that i can use tensorflow 2.0.0 which supports cuda-10.0 (this is the latest version with libcublas prepackaged)?
Correction on 7:18 "conda install -y jupyter" use this if not success using jup only...
I want to know the laptop's specs I should have to start in deep learning
I get this:
Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
[[node model/conv1d/conv1d (defined at :3) ]] [Op:__inference_distributed_function_1394]
Function call stack:
distributed_function
when i try to train one epoch with a cnn =(
Sir, please can u suggest me the current CUDA and cuDNN versions for my GPU (GTX 1060 6gb) ?
I have tried a lot.. in the end my gpu is not detected by TF ..
Hi Jeff, I am using TensorFlow version 1.0.0 with python 3.5 to train my RNN model on Anaconda spyder. But now since I started the training I found out that it maxes out my CPU. I have an NVIDIA graphics card GTX 1050 and want to use it with minimal changes in my code. How can I do this.
Hello jeff, I'm starting in this world of neuronal networks, I have a question about of this topic: Do you know if is possible to configurate like as gpu an integrate gpu how is AMD with Radeon vega 11? I configurated the enviroment wirh cpu but i want to know if is possible configurate that GPU. Thanks!
"conda install -c anaconda tensorflow-gpu" Doesn't work for windows. do you help me for making a correction? It also error in "PackagesNotFoundError: The following packages are not available from current channels:". Please help me.
try using 'pip install tensorflow' it has GPU support too.
thanks a lot
I have been having a hard time with conda installation on my PC. It don't seem to show the navigation menu anywhere and when I click on the Anaconda cmd, it opens and closes almost immediately. I've tried different suggestions but it hasn't been better, will love to hear from you also
Thank you! That was really helpful, things lot easy by this tutorial..
Thank you once again
Love from india❤️
Thanks for the vid! I had a problem where when I tried to run the
t81_558_class_01_1_overview on jupyter notebook, this error occurs: AttributeError: module 'tensorflow' has no attribute 'compat'. Id appreciate any help!
Some suggestions here on this error: github.com/tensorflow/tensorflow/issues/37525
While trying to install miniconda, I see Python 3.8 and python 2.7. In this video, it shows Python 3.7. Its a problem because I read at places that tensorflow doesn't support Python 3.8.How should I proceed?
Same as me
repo.anaconda.com/miniconda/
Once there, hit CTRL+F then type "py37" to search for 3.7 ones
hi sir,
I have a question.
when startup deploys their models on cloud, I don't think they would be running instance 24*365 it would be very costly.
can you please explain how they do it?
The cloud can get costly. My only suggestion is to use automation and use for only as much time as is needed. Also do not use more hardware than you need. Multiple GPUs won't help some models.
I wonder who dosen't know what a gpu is and actuallely needs this video :)
I have been following this and there was no DLL error. Tensorflow recognize my RTX 2070 GPU and used all of its memory but during training I open my task manager window and see that CUDA is not working
Usually I will see GPU usage show up on task manager, but not all networks will benefit from a GPU. That might be the issue.
so after install, what to do with it ?
Subscribe to my channel of course! :) Or follow along with my course... sites.wustl.edu/jeffheaton/t81-558/
great video sir, I just want to know. if I installed python ver 3.9 and im using the latest miniconda3 that use python 3.8. will it affect the miniconda if I tick the box of "add miniconda3 to my PATH environment variable".??
at 8:52 , you use python= 3.7 because tensorflow required to use that version or because based on miniconda python version u downloaded?. mine use python 3.8 for the miniconda. should I write conda create --name tensorflow python=3.8??
Thank you for your videos, looking for your help guys, I installed Python 3.8 on my Windows when I got to install tensorflow with the command " conda install -c anaconda tensorflow" an error says its incompatible, your advice please
حبيبي نزل بايثون نسخة ٣.٧ لان النسخة الي منزلها بعدها متدعم ال tensorflow
where do i write the code in the one
Hi Sir! can i use it for Linux OS
Thanks a lot Prof Heaton. This video helped me a lot.
Thanks so much for the useful video I followed everything and it appeared that tensorflow uses GPU but every time I try to continue a message appears in jupyter NB : Kernel appears to have died it will restart automatically , Also this error appears
: Attempting to fetch value instead of handling error Internal: failed to get device attribute 13 for device 0: CUDA_ERROR_UNKNOWN: unknown error
I don't know why .. so could you help me solving such problem ? thank you so much
is it fine with python 3.8 too?
How about for Macs?
Sure, ua-cam.com/video/LnzgQr14p7s/v-deo.html