NVidia Benchmarks for specific Artificial Intelligence tasks (Generative Adversarial Networks)

I started to do some benchmarks in order to plan the future needed computers and tools for Generative Adversarial Networks (GAN)

bechmark-adrian-minst-1

3 computers until now (all on Windows 10)
Elapsed Time for CPU = 5 hours, 47 minutes, 26 seconds
Elapsed Time for GPU NVidia 920M = 1 hour, 39 minutes, 15 seconds
Elapsed Time for GPU NVidia 940MX + SSD = 1 hour, 13 minutes, 16 seconds

GPU and SSD are a must so you need to
– install tensorflow-gpu instead of tensorflow
– have python 3.6 (on 3.7 id does not work yet, see tensorflow site https://www.tensorflow.org/install/pip)
– install Cuda 9 (I also installed Cuda 10 but it does not work – you can have both Cuda installed)
https://developer.nvidia.com/cuda-90-download-archive
Cuda 10 is here https://developer.nvidia.com/cuda-downloads
– install cuDNN for Cuda 9 (register at NVidia, download package and copy 3 files as described on site )
https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html

Tensorflow news:
TensorFlow 1.13.0-rc2 @aselle aselle released this 5 days ago
TensorFlow GPU binaries are now built against CUDA 10 and TensorRT 5.0.
Support for Python3.7 on all operating systems.

https://github.com/tensorflow/tensorflow/releases

I tested with Tensorflow nightly 1.13.0-rc2, Cuda 10, cuDNN v7.4.2. I had to install a new driver from Nvidia GeForce Experience application to make the system work. I am still on Python 3.6.8. I will come back here after Python update. Now I am benchmarking this new configuration

Bad news – Cuda compute capability minimum requirements (solved on Tensorflow nightly 1.13.0-rc2, Cuda 10, cuDNN v7.4.2)

I upgraded to Tensorflow 1.11.0 and the game ended for NVidia 920M (Cuda compute capability required is 3.7 > cuda compute capability of NVidia 920M). Check your board on this table https://developer.nvidia.com/cuda-gpus

Tests where done on super algorithms described at https://www.pyimagesearch.com/

I will add more NVidia boards and configurations on this article 

Enjoy Artificial Intelligence and Generative Adversarial Networks (GAN) !

Dorin

P.S. Strange errors while using other configurations

File “C:\Program Files\Python36\lib\site-packages\tensorflow\python\keras\backend_init_.py”, line 22, in
from tensorflow.python.keras.impl.keras.backend import abs

File “C:\Users\dorin\AppData\Local\Programs\Python\Python36\lib\ctypes_init
.py”, line 348, in init
self._handle = _dlopen(self._name, mode)
OSError: [WinError 126] The specified module could not be found

File “C:\Users\dorin\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\platform\self_check.py”, line 97, in preload_check
% (build_info.cudnn_dll_name, build_info.cudnn_version_number))
ImportError: Could not find ‘cudnn64_7.dll’. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. Note that installing cuDNN is a separate step from installing CUDA, and this DLL is often found in a different directory from the CUDA DLLs. You may install the necessary DLL by downloading cuDNN 7 from this URL: https://developer.nvidia.com/cudnn

……………

 File “C:\Users\dorin\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\saved_model\model_utils\export_utils.py”, line 51, in <module>
    ModeKeys.TRAIN: signature_constants.DEFAULT_TRAIN_SIGNATURE_DEF_KEY,
AttributeError: module ‘tensorflow.python.saved_model.signature_constants’ has no attribute ‘DEFAULT_TRAIN_SIGNATURE_DEF_KEY’
……………….
Ignoring visible gpu device (device: 0, name: GeForce 920M, pci bus id: 0000:08:00.0, compute capability: 3.5) with Cuda compute capability 3.5. The minimum required Cuda capability is 3.7.

 

A heart that can control a musical instrument – Pink Floyd

I found one “Michael Jackson – Paganini” similar exquisite talent inside The Pink Floyd world. Joe Coker was born with an orchestra inside his voice, and he worked all his life to connect his heart with his voice. Other people had to reinvent themselves to be the gods of the violin, dance, or guitar.
5 performances of Serin Halim on Pink Floyd tunes on the playlist below.
He is body and soul dedicated to music. I met him several times and it reminds me of Santana (“What is my music offer to the day and Humanity”)
Serin completely changes while he is singing, getting to a completely new world like Alice in Wonderland . Probably, Tony Robbins will call it double personality. I call it Paganini / Michael Jackson effect = getting first to a Music Universe and singing from there

Roger Waters is starting at the 6th tune (Money is singed by Roger Waters with Donald Trump voice at the beginning)

Also a different playlist with The Eagles and Led Zepelin covers are here

[Reiland Olivier comment]
The guy was absolutely brilliant ! It’s so funny as i was just standing at that exact time on your right side !!
At the end an italian cop ordered it to stop his playing . I was so hangry against this cop .. 🙁 Thanks for this video, mate !
[Reply]
That day we made History ! Well … Pink Floyd with their genius, Serin with his love and talent, and we come last but not least because we had another needed quality: We do not move when we hear real music and feelings. We know that Pantheon was build for this kind of art, too.

Enjoy !

Dorin

P.S. Unfortunately, Pantheon is closed for street musicians starting august 2018. Read the italian articles below
Roma-la-musica-degli-artisti-di-strada-vietata-in-centro

S.O.S. Artisti di Strada

Is the sound so dangerous ?  If is hard to measure by the police squads so a “no sound” politics was approved by The Mayor Office for the Center of Rome – sound weapon