{"author":"George Hotz","author_email":"72895+geohot@users.noreply.github.com","author_time":1708012539,"commit_time":1708012539,"committer":"GitHub","committer_email":"noreply@github.com","hash":"b1c0d8c99db731a210cae035f052fe72ff6391a3","message":"remove cpu and torch backends (#3399)\n\n* remove cpu and torch backends\r\n\r\n* don't copy to cpu\r\n\r\n* use clang instead of cpu\r\n\r\n* multitensor gathers on the first device\r\n\r\n* clang is cpu + use default\r\n\r\n* fixup\r\n\r\n* bugfix","parents":["75f7e21a8010e953aa304798adf21ad29030a025"],"tree_hash":"8a121cf6a59fa37c023877d66ce4afc61c1bf316"}