George Hotz e1d13bc4fe add GGUF IQ4_XS support (#15766) ...

* add GGUF IQ4_XS support

* gguf 21

* gguf 21

* use plus

* ggml_common autogen for constant arrays

* fix

* ggml_common in autogen

* inline

2026-04-17 14:43:39 +08:00

..

codegen

remove opts arg in get_program, everything uses opts_to_apply [pr] (#15767)

2026-04-16 22:42:43 +03:00

engine

jit: capturedjit is linear (#15743)

2026-04-16 14:54:39 +03:00

llm

refactor llm into files (#15780)

2026-04-17 12:33:11 +08:00

mixin

softmax and friends to mixin (#15778)

2026-04-16 23:03:37 -04:00

nn

add GGUF IQ4_XS support (#15766)

2026-04-17 14:43:39 +08:00

renderer

amd: EMU DPP support (#15719)

2026-04-14 14:58:41 +08:00

runtime

add GGUF IQ4_XS support (#15766)

2026-04-17 14:43:39 +08:00

schedule

another form of assign to itself (#15770)

2026-04-16 15:17:19 -04:00

uop

softmax and friends to mixin (#15778)

2026-04-16 23:03:37 -04:00

viz

CALL with return value is FUNCTION (#15758)

2026-04-16 13:25:07 +08:00

__init__.py

start function and add walk rewrite (#14992)

2026-02-25 13:56:27 +08:00

callify.py

CALL with return value is FUNCTION (#15758)

2026-04-16 13:25:07 +08:00

device.py

DEV: suggest alternatives (#15732)

2026-04-14 23:42:32 -04:00

dtype.py

print function timing with DEBUG=2 (#15468)

2026-03-25 19:07:32 +08:00

function.py

delete old UOp.size [pr] (#15756)

2026-04-15 23:21:00 -04:00

gradient.py

CALL with return value is FUNCTION (#15758)

2026-04-16 13:25:07 +08:00

helpers.py

add NO_COLOR (#15765)

2026-04-16 22:44:55 +03:00

py.typed

add a single py.typed (#6083)

2024-08-14 17:31:46 -07:00

tensor.py

softmax and friends to mixin (#15778)

2026-04-16 23:03:37 -04:00