The software is of good quality. The images are high-resolution and closely resemble the reference model, with no particular artefacts on the feet or hands (except in a few cases), but it has some significant limitations which I hope will be rectified as soon as possible:
- The images created are displayed without the prompt and the settings used to generate them. This makes it impossible to recreate similar images unless you have copied the prompt and parameters elsewhere (1 star)
- Only six photos were used to train the model, one of which showed a smile. In 90 per cent of cases, the generated images always featured the same facial expression (the smile) and the same gaze (2 stars). Training the model on a larger number of photos with a wider range of facial expressions would certainly have improved the quality of the images produced.
- Since I purchased the software, my Appsumo credits have never been updated automatically each month. I have always had to send a message via the software’s messaging system (chat) and wait 2–3 days for the credits to be updated.
I hope the software company takes these points on board and implements these features as soon as possible.
The quality and likeness of the generated photos are very good, but these shortcomings are significant.