The domestication of many crops resulted in reduced genetic diversity. One of the most extreme examples is in cultivated sunflowers, Helianthus annuus, for their oils and confectionery cultivars. Due to these breeding practices, only a certain number of sunflower species were selected to emphasize these favorable traits. Doing so creates a selective sweep, where a mutation fixes in a population and then reduces the genetic variation that are similar to the mutation. This signifies extremely low genetic diversity compared to the wildtype and causes a population bottleneck effect. Understanding whether a specific trait or a variety of similar traits are more popular among an entire population is still a current topic being studied. Using machine learning and simulation tools, we investigate selective sweeps using a deep learning approach to identify hard and soft sweeps related to the population bottleneck effect. Hard sweeps indicate that there is a single variant that will arise in frequency over time, while soft sweeps indicate multiple alleles that will arise over time. The main difference between hard sweep and soft sweep is that the soft sweep can identify other variants that are not strong of a match within that population. These sweeps were created through a highly computerized cluster by a training dataset program called diploSHIC. In our findings, we conclude that there are multiple variants of confectionary cultivars that are found among different species of sunflowers in both hard sweeps and soft sweeps. With our findings, we detail the relationship between the genetic architecture of phenotypic traits in cultivated sunflowers in the context of potential selective sweeps.