osanseviero's picture
Add image-to-image tag for discoverability
8de1bff
|
raw
history blame
541 Bytes
metadata
tags:
  - fastai
  - pytorch
  - image-to-image

Model description

This repo contains the trained model for Style transfer using vgg16 as the backbone.

Full credits go to Nhu Hoang

Motivation: Style transfer is an interesting task with an amazing outcome.

Training and evaluation data

Training hyperparameters

The following hyperparameters were used during training:

Hyperparameters Value
name Adam
learning_rate 3e-5
training_precision float16