George HotzandGitHub 9d05dfb6f4 move JIT graphing into CapturedJit (#5852) ...

* move JIT graphing into CapturedJit

* better

* _jit_cache

* clear inputs cleanup

* test_pickle_jit with graph + cleanup

* 0 is fine to start

* support None in bufs

* alloc real buffers

* cleaner

2024-07-31 20:48:17 -07:00

..

codegen

diff symbolic with uops [run_process_replay] (#5841)

2024-07-31 15:15:01 -07:00

engine

move JIT graphing into CapturedJit (#5852)

2024-07-31 20:48:17 -07:00

nn

uniform init to match torch (#5494)

2024-07-15 12:07:44 -07:00

renderer

work to make GEMV fast (#5824)

2024-07-30 17:41:40 -07:00

runtime

hcq fix timestamp around kernel (#5837)

2024-07-31 16:12:27 +03:00

shape

unify negative shape creation to raise ValueError (#5817)

2024-07-30 13:42:59 -04:00

__init__.py

add mnist tutorial

2024-04-25 16:08:32 +08:00

device.py

use decimal for timestamps for more precision [run_process_replay] (#5823)

2024-07-30 15:06:14 -07:00

dtype.py

pretty print lazy op per default (#5505)

2024-07-18 09:34:08 -07:00

function.py

threefry as pattern matcher (#5371)

2024-07-13 11:59:03 -07:00

helpers.py

move GraphException to jit.py (#5744)

2024-07-26 19:01:12 -04:00

lazy.py

LazyBuffer.const type check cleanup [run_process_replay] (#5795)

2024-07-29 16:17:14 -04:00

multi.py

add RING_ALLREDUCE_THRESHOLD (#5835)

2024-07-31 16:13:09 +03:00

ops.py

add test to validate lazyops dims (#5845)

2024-07-31 12:59:38 -07:00

tensor.py

allow specify splits in shard, handle multiple different splits in MLB.e (#5599)

2024-07-30 19:33:04 -07:00