This study evaluates a Transformer model trained for binary classification to determine the presence or absence of crops in Jalisco, Mexico. Using NASA Harvest's Openmapflow library and the Geowiki Landcover 2017 dataset, the model leverages remote sensing data to produce accurate inference maps of satellite images. The Transformer model, based on the TSAI library, utilized features from all bands of Sentinel-1 and Sentinel-2 satellites, ERA5 and SRTM of each pixel during the inference process. After 15 epochs, the model demonstrated stability and achieved an accuracy of 88.31%. Applied to a regional scale crop classification near the Guadalajara Metropolitan Zone, the model showed promising distinctions between cultivated and non-cultivated areas. Building upon NASA Harvest's Openmapflow initiative, this research establishes a fundamental step towards refining the model with local datasets, thereby improving accuracy in crop mapping and land use analysis. The findings highlight the potential of this approach to serve as a foundational basis for developing a real-time crop monitoring program, specifically focusing on a broader classification range of agricultural classes in the region.