chenyu 94926d00d8 fix rand > uint32.max (#15330) ...

need to keep low and high as 1D tensor.
`PYTHONPATH=. LLAMA3_SIZE=405B python3 examples/mlperf/models/flat_llama.py` works now

2026-03-17 22:00:01 -04:00

..

apps

llm: fix qwen3 moe topk renormalization (#15201)

2026-03-17 12:57:33 +08:00

codegen

update no_vectorized_index [pr] (#15313)

2026-03-17 03:05:23 -04:00

engine

precompile backward (#15311)

2026-03-17 15:28:40 +08:00

mixin

move more to mixins (#14780)

2026-02-16 11:35:00 +08:00

nn

embedded bwd vocab shard (#15001)

2026-03-16 19:37:16 -07:00

renderer

sqtt: add CDNA ops enum, show in viz (#15140)

2026-03-17 09:38:42 +09:00

runtime

system: remote (#15318)

2026-03-17 19:25:37 +08:00

schedule

update fix_store_after_hazard (#15309)

2026-03-16 23:55:59 -04:00

uop

disallow reshape from None shape [pr] (#15322)

2026-03-17 05:46:53 -04:00

viz

viz: rename to Start Cycle for the sqtt graph (#15320)

2026-03-17 18:53:06 +09:00

__init__.py

start function and add walk rewrite (#14992)

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

device.py

remove dmaref (#15321)

2026-03-17 17:52:09 +08:00

dtype.py

images with height=1 have less strict width rules (#15325)

2026-03-17 07:07:22 -04:00

function.py

set allow_implicit=False by default (#15319)

2026-03-17 17:14:38 +08:00

gradient.py

precompile backward (#15311)

2026-03-17 15:28:40 +08:00

helpers.py

system: print output when err (#15230)

2026-03-12 18:30:49 +08:00

py.typed

add a single py.typed (#6083)

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

tensor.py

fix rand > uint32.max (#15330)

2026-03-17 22:00:01 -04:00