qazal be09cc87c1 Bitcast support / fast bf16 load (#2011) ...

* bitcast renderers

* fast llama load

* make it one kernel

* regression testing p1: re-enable test_dtype for all backends

fix GPU

* regression testing p2: fuzz all possible cases against numpy

remove hancoded tests since the fuzzer covers them

* define ushort

* fix indent, probably need flake8 back for CI to catch

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>

2023-12-05 16:19:28 -08:00

..

triton.py

Bitcast support / fast bf16 load (#2011)

2023-12-05 16:19:28 -08:00