The Artwork
alignDRAW is a collection of 2709 images, generated from 36 unique text prompts, that were created in 2015.
Paper Prompts
In the published academic paper, ‘Generating Images from Captions with Attention‘, 21 unique text prompts, each producing a set of 8 images, were used to illustrate the capabilities of the alignDRAW model. These 168 individual images are categorised as ‘Paper Prompts’.
Each image is 32 x 32 pixels. The NFT editions are minted fully on-chain from their original files (more details here) and the print editions scaled using ‘nearest neighbour’ to faithfully preserve the original pixelation.
Images of the print editions are shown below from Fellowship’s booth at Paris Photo. 2023. For images of the digital editions and release information see Project Details 
Process Prompts
Mansimov also saved alignDRAW outputs from a further 21 text prompts, each producing a set of 121 images. These 2,541 images, created by the same alignDRAW model, are categorised as ‘Process Prompts’.
Each image is 32 x 32 pixels. The NFT editions are minted fully on-chain from their original files (more details here) and the print editions scaled using ‘nearest neighbour’ to faithfully preserve the original pixelation.
Images of the print editions are shown below from Fellowship’s booth at Paris Photo. 2023. For images of the digital editions and release information see Project Details