Network Security Internet Technology Development Database Servers Mobile Phone Android Software Apple Software Computer Software News IT Information

In addition to Weibo, there is also WeChat

Please pay attention

WeChat public account

Shulou

Let tersorflow fly with GPU.-MX150

2025-04-10 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Servers >

Share

Shulou(Shulou.com)06/02 Report--

I think it is an indisputable fact to use GPU instead of CPU in in-depth learning and training. Although MX150 is not listed on the official website below, MX150 actually supports CUDA.

Https://developer.nvidia.com/cuda-gpus

Environment

Millet Notebook Air 13

OS: win10CPU: I7 7500UGPU: MX150GPU driver: 425.25tensorflow: 1.13.1tensorflow-gpu: 1.13.1visual studio: 2019 install TensorFlow

Now the whl file of TensorFlow is well packaged and can basically be installed, but if there is no driver such as CUDA, there will be an error when importing tensorflow.

Pip install tensorflow tensorflow-gpu installation dependency

In order to install Nvidia's CUDA suite, you also need to install visual studio, because the relevant compilation environment of windows is tied to visual studio, even if you only want to install part of it, you still have to install visual studio.

Install visual studio

The windows installer won't be too difficult, it's the next step, the next step.

Install cuda toolkits

CUDA toolkits 10.0

Https://developer.nvidia.com/cuda-zone

It can be installed by default. Next step.

Install cudnn

Cudnn 7.6.0.64

Https://developer.nvidia.com/cudnn

Be careful not to be greedy for the latest version, tensorflow official may not support it!

After decompressing cudnn, put it in C:\ tools\ cuda

Finally, add the environment variable:

C:\ tools\ cuda\ binC:\ Program Files\ NVIDIA GPU Computing Toolkit\ CUDA\ v10.1\ extras\ CUPTI\ lib64

The official statement is as follows:

Software requirements

The following NVIDIA ®software must be installed on the system:

NVIDIA ®GPU driver-CUDA 10.0 requires 410.x or later.

CUDA ®Toolkit-TensorFlow supports CUDA 10.0 (TensorFlow 1.13.0 and later)

The CUPTI that comes with the CUDA toolkit.

CuDNN SDK (7.4.1 and later)

(optional)

TensorRT 5.0to reduce latency and improve throughput for inference on some models.

Reference page:

Https://www.tensorflow.org/install/gpu

Compatibility testing of each version

Https://www.tensorflow.org/install/source#linux

Installation reference:

Https://towardsdatascience.com/installing-tensorflow-with-cuda-cudnn-and-gpu-support-on-windows-10-60693e46e781

Https://medium.com/@johnnyliao/%E5%9C%A8nvidia-mx150%E7%9A%84win10%E5%AE%89%E8%A3%9Dcuda-toolkit-cudnn-python-anaconda-and-tensorflow-91d4c447b60e

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.

Views: 0

*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.

Share To

Servers

Wechat

© 2024 shulou.com SLNews company. All rights reserved.

12
Report