Making AI Art with Midjourney

Midjourney is an independent research lab, exploring AI applications, and recently they launched a bot capable of producing art with text inputs.

<p dir="auto">Others were able to produce AI models to generate images from text, such as: <a href="https://openai.com/dall-e-2/" target="_blank" rel="nofollow noreferrer noopener" title="This link will take you away from hive.blog" class="external_link">DALL-E and <a href="https://app.wombo.art/" target="_blank" rel="nofollow noreferrer noopener" title="This link will take you away from hive.blog" class="external_link">Wombo. But until know, Midjourney was the best experience, at least for me. <h2>My AI paintings 🎨 <p dir="auto">I will explain you ahead, how to use the Midjourney algorithm on discord, but first check my paintings generated out of text prompts! <p dir="auto"><center><em>"Data Scientist walking around a dystopian future market and buying goods with cryptocurrency" <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKDgaR3LwFLJjf2FB1q1NqXLJYLjdyB8pWbSR5MVpS8UdpkyByLgSDKwjKCxj6c.png" alt="Data Scientist walking around a dystopian future market and buying goods with cryptocurrency.png" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKDgaR3LwFLJjf2FB1q1NqXLJYLjdyB8pWbSR5MVpS8UdpkyByLgSDKwjKCxj6c.png 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKDgaR3LwFLJjf2FB1q1NqXLJYLjdyB8pWbSR5MVpS8UdpkyByLgSDKwjKCxj6c.png 2x" /> <p dir="auto"><center><em>"A leek with the face of wolverine, trying to fight a rainbow of puppets" <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AK3evgdcjGSkqxw7RPfk2mttG8C9mWFhasVMZdNuwgdv9qYun5zfdqCzWSNmoYi.jpg" alt="A leek with the face of wolverine, trying to fight a rainbow of puppets.jpg" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AK3evgdcjGSkqxw7RPfk2mttG8C9mWFhasVMZdNuwgdv9qYun5zfdqCzWSNmoYi.jpg 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AK3evgdcjGSkqxw7RPfk2mttG8C9mWFhasVMZdNuwgdv9qYun5zfdqCzWSNmoYi.jpg 2x" /> <p dir="auto"><center><em>"Fernando Pessoa in a very futuristic style" <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AK7wyTdBhrae8iKDSzTecpJWmoJBmBDbLrGBRhgRi2Y9aHd4rYhwmPu4wHaVZZv.png" alt="mac.rodrigues_Fernando_Pessoa_in_a_very_futuristic_style_5640d2ee-97e6-473f-8e56-448cf8aa6705.png" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AK7wyTdBhrae8iKDSzTecpJWmoJBmBDbLrGBRhgRi2Y9aHd4rYhwmPu4wHaVZZv.png 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AK7wyTdBhrae8iKDSzTecpJWmoJBmBDbLrGBRhgRi2Y9aHd4rYhwmPu4wHaVZZv.png 2x" /> <p dir="auto"><center><em>"Gorgeous Mexican woman with freckles, dark hair, dark eyes and chiapas clothes" <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKbEkFDiFtgMNvLPJk8U5jsogADDiKp3XuE68k3wMeuU6Mu5vzN5qgy4MrPUSgg.png" alt="gorgeous mexican woman with freckles, dark hair, dark eyes and chiapas clothes.png" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKbEkFDiFtgMNvLPJk8U5jsogADDiKp3XuE68k3wMeuU6Mu5vzN5qgy4MrPUSgg.png 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKbEkFDiFtgMNvLPJk8U5jsogADDiKp3XuE68k3wMeuU6Mu5vzN5qgy4MrPUSgg.png 2x" /> <h2>Using Midjourney <p dir="auto">The AI bot it's still on beta version, but you can try it on discord. You can access their discord through <strong><a href="https://www.midjourney.com/home/" target="_blank" rel="nofollow noreferrer noopener" title="This link will take you away from hive.blog" class="external_link">Midjourney's website. <p dir="auto">Once on Midjourney's Discord, you have the instructions to use it, basically you just need to join a "newbie" channel and prompt: <pre><code>/imagine <your quote> <p dir="auto">The output will be 4 images like this: <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKKSobQqrUT6555YtbLgBu2UUE6zrLhoZFviuGUwZ7rzGHPDUsDnPtx5qa8BoBB.png" alt="mac.rodrigues_Fernando_Pessoa_in_a_very_futuristic_style_9ffbd501-c7d0-48e8-a8b9-0fe1e9f42224.png" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKKSobQqrUT6555YtbLgBu2UUE6zrLhoZFviuGUwZ7rzGHPDUsDnPtx5qa8BoBB.png 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/AKKSobQqrUT6555YtbLgBu2UUE6zrLhoZFviuGUwZ7rzGHPDUsDnPtx5qa8BoBB.png 2x" /> <p dir="auto">You can then decide to get an high resolution image from the 4, or get more versions of one of the images. <p dir="auto">Note that you don't have many tries as a free user (around 25 queries), so make sure to prompt the best ideas that come to your mind! (or not, just have fun 😄) <h2>Behind the hood 💻 <p dir="auto">The technical background behind the magic easily gets very complex to explain, but mainly includes 3 steps to output the image: The <strong>training data, the <strong>deep learning model, and the <strong>latent space. <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/23t79FbJnfPssyiXd3qX4tYCGxayko8tEjTnx9MmSp3nXgmVDpXDdGQ8fW6x249Y6tv3f.png" alt="Schematic AI ART.png" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/23t79FbJnfPssyiXd3qX4tYCGxayko8tEjTnx9MmSp3nXgmVDpXDdGQ8fW6x249Y6tv3f.png 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/23t79FbJnfPssyiXd3qX4tYCGxayko8tEjTnx9MmSp3nXgmVDpXDdGQ8fW6x249Y6tv3f.png 2x" /> <p dir="auto">The training data passes through the deep learning model (CNN) and needs to be encoded into a low-dimensional latent space before classification. The later basically allows for data compression. <p dir="auto">In the end of the training, the last layer of the model captures the patterns of the input that are needed for the image classification. In the latent space, images that are labelled as the same object have very close representations. <p dir="auto"><img src="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/23x1Yjd8Rks2k3w4x8s8gk6tX1AgTFawgaiyMDamgLy5WGEMRAwPYr4ZNirGrbtXA9VAR.png" alt="Source: https://www.baeldung.com/cs/dl-latent-space" srcset="https://images.hive.blog/768x0/https://files.peakd.com/file/peakd-hive/macrodrigues/23x1Yjd8Rks2k3w4x8s8gk6tX1AgTFawgaiyMDamgLy5WGEMRAwPYr4ZNirGrbtXA9VAR.png 1x, https://images.hive.blog/1536x0/https://files.peakd.com/file/peakd-hive/macrodrigues/23x1Yjd8Rks2k3w4x8s8gk6tX1AgTFawgaiyMDamgLy5WGEMRAwPYr4ZNirGrbtXA9VAR.png 2x" /> <p dir="auto">For more technical information, I found this video very enlightening: <p dir="auto"> <p dir="auto">Other sources: <ul> <li><span><a href="https://www.baeldung.com/cs/dl-latent-space" target="_blank" rel="nofollow noreferrer noopener" title="This link will take you away from hive.blog" class="external_link">https://www.baeldung.com/cs/dl-latent-space <li><span><a href="https://towardsdatascience.com/understanding-latent-space-in-machine-learning-de5a7c687d8d" target="_blank" rel="nofollow noreferrer noopener" title="This link will take you away from hive.blog" class="external_link">https://towardsdatascience.com/understanding-latent-space-in-machine-learning-de5a7c687d8d