Using Deep Neural Networks for Image Reconstruction
سه سال پیش منتشر شده
تعداد بازدید: 434
کد پروژه: 366082
شرح پروژه
You are going to implement a deep neural network to predict the missing part of an image using the PyTorch framework. The input is the 14x28 upper half image of a handwritten digit and the network should learn to generate as output the corresponding 14x28 lower half of the image.
MNIST dataset should be used.
Implement an encoder-decoder network where the encoder takes the 14x28 upper half, passes it through one or more convolutional layers followed by one or more fully-connected layers to generate a code, and then the decoder does the inverse operation using one or more fully-connected layers followed by one or more deconvolutional layers. The aim is to achieve small test reconstruction loss
with a small network.
1. Explain your architecture,
2. Include the figure that plots the convergence of training/test loss during training,
3. Report the average training and test loss values at the end of training,
4. Report the number of parameters,
5. For the first 16 test instances, show the target 28x28 image, the 14x28 input image, and the predicted 14x28 image as shown in the attached image.
این پروژه شامل 1 فایل مهم است، لطفا قبل از ارسال پیشنهاد حتما نسبت به بررسی این فایل اقدام فرمایید.
مهارت ها و تخصص های مورد نیاز
مبلغ پروژه
900,000 تومان
مهلت برای انجام
3روز
وضعیت مناقصه
انجام شده
درباره کارفرما
عضویت سه سال پیش
