George HotzandGitHub 5cac6338a4 apply the multitensor optimizations in lazy.py (#2901) ...

* apply the multitensor optimizations in lazy.py

* less lines

* hack for webgpu

* save a line

2023-12-21 13:55:49 -08:00

..

codegen

Fix exponential behavior in lazyops (#2890)

2023-12-20 22:06:50 -08:00

features

new lazy, benchmark (#2878)

2023-12-20 14:33:21 -08:00

nn

remove realize from optimizer (#2880)

2023-12-20 16:42:41 -08:00

renderer

render_locals takes a dtype (#2873)

2023-12-21 14:15:28 -05:00

runtime

use np.less and torch.lt for CMPLT (#2899)

2023-12-21 14:37:24 -05:00

shape

apply the multitensor optimizations in lazy.py (#2901)

2023-12-21 13:55:49 -08:00

__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

Fix int child count (#2882)

2023-12-20 21:06:27 -08:00

helpers.py

move graph to runtime, check line count with sz.py (#2842)

2023-12-18 20:30:06 -08:00

jit.py

new lazy, benchmark (#2878)

2023-12-20 14:33:21 -08:00

lazy.py

apply the multitensor optimizations in lazy.py (#2901)

2023-12-21 13:55:49 -08:00

mlops.py

Move the webgpu CMPLT hack to one place (#2895)

2023-12-21 11:14:56 -05:00

ops.py

Move the webgpu CMPLT hack to one place (#2895)

2023-12-21 11:14:56 -05:00

realize.py

get_lazyops() -> lazyops (#2884)

2023-12-20 18:04:49 -08:00

tensor.py

don't match dtype with first input in where (#2898)

2023-12-21 13:02:15 -05:00