chenyu c0f76ed4ea transformer kvcache and mask have same dtype as input (#2771) ...

* transformer kvcache and mask have same dtype as input

* don't use `=0` in cstyle ternary where

* (bool)

* where float16 test

2023-12-14 22:41:51 -05:00

..

codegen

remove duplicated dtype in DEFINE_GLOBAL args (#2768)

2023-12-14 15:42:36 -05:00

features

More simplification in to_image_idx and symbolic (#2679)

2023-12-13 12:30:44 -05:00

nn

changes to prep for new lazy (#2748)

2023-12-13 10:28:22 -08:00

renderer

transformer kvcache and mask have same dtype as input (#2771)

2023-12-14 22:41:51 -05:00

runtime

set default str for CStyleLanguage.arg_int_prefix (#2756)

2023-12-13 20:23:27 -05:00

shape

More simplification in to_image_idx and symbolic (#2679)

2023-12-13 12:30:44 -05:00

__init__.py

move device to device.py (#2466)

2023-11-27 11:34:37 -08:00

device.py

Fix double exception in __del__() when __init__() raises exception. (#2738)

2023-12-13 15:46:11 -08:00

graph.py

ruff checks the max line length is 150 (#2734)

2023-12-12 17:34:47 -08:00

helpers.py

remove _arg_int32 internal type (#2767)

2023-12-14 14:17:14 -05:00

jit.py

ruff checks the max line length is 150 (#2734)

2023-12-12 17:34:47 -08:00

lazy.py

ruff checks the max line length is 150 (#2734)

2023-12-12 17:34:47 -08:00

mlops.py

some unary functions cast int input into float (#2740)

2023-12-13 00:10:29 -05:00

ops.py

changes to prep for new lazy (#2748)

2023-12-13 10:28:22 -08:00

realize.py

ruff checks the max line length is 150 (#2734)

2023-12-12 17:34:47 -08:00

tensor.py

arange default dtype to int and zeros/ones default to float (#2769)

2023-12-14 17:53:00 -05:00