To generate the paintings I used GPT-2, a linguistic neural model able to generate appropriate text from a given incipit, and AttnGAN, a Generative Adversarial Network that converts words and sentences in synthetic images.

 

The first one was employed to create the portrait from the title, and then GPT would be used again to explain said portrait after a short description of what I saw in it.

 

Lastly I got the portraits’ descriptions, in mp3, through Polly, an AI which belongs to Amazon Web Services that is dedicated to vocal synthesis.

 

To generate the paintings I used GPT-2, a linguistic neural model able to generate appropriate text from a given incipit, and AttnGAN, a Generative Adversarial Network that converts words and sentences in synthetic images.

 

The first one was employed to create the portrait from the title, and then GPT would be used again to explain said portrait after a short description of what I saw in it.

 

Lastly I got the portraits’ descriptions, in mp3, through Polly, an AI which belongs to Amazon Web Services that is dedicated to vocal synthesis.

 

To generate the paintings I used GPT-2, a linguistic neural model able to generate appropriate text from a given incipit, and AttnGAN, a Generative Adversarial Network that converts words and sentences in synthetic images.

The first one was employed to create the portrait from the title, and then GPT would be used again to explain said portrait after a short description of what I saw in it.

Lastly I got the portraits’ descriptions, in mp3, through Polly, an AI which belongs to Amazon Web Services that is dedicated to vocal synthesis.

To generate the paintings I used GPT-2, a linguistic neural model able to generate appropriate text from a given incipit, and AttnGAN, a Generative Adversarial Network that converts words and sentences in synthetic images.

The first one was employed to create the portrait from the title, and then GPT would be used again to explain said portrait after a short description of what I saw in it.

Lastly I got the portraits’ descriptions, in mp3, through Polly, an AI which belongs to Amazon Web Services that is dedicated to vocal synthesis.