Carrying out AI and item recognition to sort recyclables is complex and would require an embedded chip able to handling these features with high performance.  We depict video clips and pictures as collections of lesser models of knowledge termed patches, each of which is akin to the toke