Train Your First GAN in Tensorflow| Complete Tutorial in Python|
Ғылым және технология
In this video, I give a complete guide to training your own generative adversarial network in python. I cover the following concepts:
1. Building Generator and Discriminator Network in Python
2. How to create custom training loop and loss functions for your GAN deep learning model.
3. How to finally generate realistic looking images using DCGAN or Deep Convolutional GAN .
4. We cover the MNIST dataset to generate realistic hand written digits in this tutorial.
Original Notebook by Tensorflow: www.tensorflow.org/tutorials/...
Do like this video and subscribe to this channel if this video helped you!
Пікірлер: 17
Dude this is the best implementation video of GANs thank you so much !!!!
Fucking hell!! this is the first video I watched on GAN implementation, it went straight through my brain.
Hey man, Thanks for the wonderful video. How can I optimize the training of this or similar codes using TensorFlow 2 on remote GPU with Ubuntu 20.0? 20.04.4 LTS (GNU/Linux 5.13.0-52-generic x86_64) NVIDIA-SMI 515.48.07 Driver Version: 515.48.07 with four NVIDIA GeForce RTX 3080 GPUs 10 GB each
Thank you. pls how can I add my own dataset instead of using the mnist one?
How can plot loss and accuracy curve to gan ? I need metric to know how the gan develops in training
I want to know that the image that we are getting at the end are the ones which are passed through the discriminatior or they are what images genrator has made regardless of they are fake or real
Does it work for text too ? like for the AI learning how to write like a person ?
I want to get which images were rejected by the discriminator during whole process how can i do that??
bro how can i load my own dataset?
Hello Sir, I have one doubt. Can GAN's models only accept images data? Can we use csv or excel file?
@eduardorosentreter
Жыл бұрын
en general, el genera un vector aleatorio, ese vector lo tienes que transformar a las dimensiones que desees de salida, su usas un excel, supongo que ese excel contenga numeros en las cuadriculas, la idea es que el tipo de dato que le pases, va a ser procesado por un discriminador previamente entrenado, el disciminador dice si esta bien o mal, ahora, ese discriminador, lo entrenas con tus datos de excel, tiene que tener como entrada el dato de excel y como salida lo que se supone que corresponda, luego de eso, entrenas a tu gan, le pasas tanto archivos excel, como los que genera tu generador. solo tienes que generar los datos de una manera que los pueda leer el discriminador A modo de Resumen: Necesitas un generador previamente entrenado que el input sea un archivo csv Luego necesitas crear el modelo de Generador, que va a tener como input, un vector aleatorio, ese vector aleatorio, la red debe encargarse de redimensionarlo y convertirlo en una variable tipo ndarray en el mismo formato que tu discriminador recibe sus input Por ultimo se crea la GAN que es la encargada de realizar el entrenamiento con los 2 modelos previamente creados.
Hey, any advice to me i am 18 confused between going to private College that's i think nothing valuable or i do online courses like digital marketing, data scientist or business analyst and find my interest in this 3 years to do that for rest of my life🙄
@alapparate8768
2 жыл бұрын
there is nothing that you can learn so it will last till rest of your life, you have to keep updating yourself. About your interests, explore them yourself what pleases you the most and go with it.
provide notebook also. or this video is not really helpfull for me
DOES THIS VIDEO EXPLAIN ON HOW TO RECONSTRUCT A 3D CT ?
@chrisidema
Жыл бұрын
no
I can't get past the accent and the audio. If you spend a bit on your audio and learn about how to do great audio, you will probably grow your channel. Controlling that deep accent would help too.