qazal 12996d3a7d green linearizer asserts for ops (#2800) ... * these asserts should pass * fix that assert * ALU dtypes * acc dtype for group_for_reduce * cast image ALUs to the base dtype * remove all casts from linearizer * fix argmax * fix multinomial * fix __getitem__ * Revert "fix __getitem__" This reverts commit 62ad719bfa. * fix MemBuffer outputs being wrong when there is an arange + ALU with a different dtype eg. fancy slicing (int, float), bert embeddings (int, long) this should be fixed in lazy instead of having to break the kernel * cleanup argmax fix * fix matmul in ints cast in the end * fix llama * skip wrong hardcoded asts in the worlds dataset * fix llama p2 * cleanup missing parts of the diff --------- Co-authored-by: George Hotz <geohot@gmail.com>		2023-12-25 10:41:54 -05:00
..
extract_policynet.py	merge kernel and optimizer (#2200)	2023-11-01 15:20:01 -07:00
extract_sa_pairs.py	merge kernel and optimizer (#2200)	2023-11-01 15:20:01 -07:00
generate_dataset.sh	op logger + replay (#2021)	2023-10-08 15:10:18 -07:00
get_action_space.py	merge kernel and optimizer (#2200)	2023-11-01 15:20:01 -07:00
helpers.py	green linearizer asserts for ops (#2800)	2023-12-25 10:41:54 -05:00
pretrain_valuenet.py	merge kernel and optimizer (#2200)	2023-11-01 15:20:01 -07:00
rl.py	fixed imports (#2185)	2023-10-30 22:07:17 -07:00
run_qnet.py	Q network for pruning BEAM / uops deduping / BEAM_ESTIMATE (#2142)	2023-10-27 10:53:06 -10:00
test_beam_search.py	autopad shapetracker for BEAM (#2375)	2023-11-22 21:05:25 -05:00
test_net.py	fixed imports (#2185)	2023-10-30 22:07:17 -07:00
test_time_linearizer.py	KOPT is over, BEAM is upstream (#2071)	2023-10-16 09:46:03 -07:00