Commit Graph

7802 Commits

Author SHA1 Message Date
Francis Lata
c84d05ee93 Merge branch 'master' into retinanet_mlperf 2025-01-29 13:25:49 -08:00
Sieds Lykles
78c0455c7a Better stable sigmoid (#8806)
Uses `1/(x*x) -> 1/x * 1/x`  together with `x/(1+x) -> 1-1/(1+x)` to
rewrite sigmoid instead of `x/((x+1)(x+1)) -> 1/(x+1)*(1-1/(x+1))`

Co-authored-by: chenyu <chenyu@fastmail.com>
2025-01-29 16:08:53 -05:00
chenyu
cac2b4e8b6 bump db version (#8805)
OptOp spec changed
2025-01-29 15:34:43 -05:00
nimlgen
a2d55fb644 am: unset CGCG override (#8804) 2025-01-29 22:28:06 +03:00
chenyu
7f606fbde4 remove DEBUG=5 in windows ci test [pr] (#8803)
DEBUG=5 prints a lot of info that's slow, and is not visible if test passed on CI.
also skip two tests that took 3 minutes in python backend
2025-01-29 14:18:17 -05:00
Ignacio Sica
260df1a17f tc_select noop (#8801)
* tc_select noop

* revert changes in test
2025-01-29 13:53:23 -05:00
FICTURE7
ec120ce6b9 Fix allocator memory alignment (#8800)
* Fix allocator memory alignment

* Run `test_ops.py` using LLVM and CLANG on Windows
2025-01-29 21:03:17 +03:00
Francis Lata
0a02a55430 Merge branch 'master' into retinanet_mlperf 2025-01-29 10:01:30 -08:00
Francis Lata
335d11281c add multi device support to retinanet eval 2025-01-29 10:00:46 -08:00
nimlgen
50ba2bb642 am: move ring to host mem (#8802) 2025-01-29 20:56:11 +03:00
chenyu
c7ca7959e6 set DISABLE_DROPOUT=1 in bert script for now (#8799) 2025-01-29 10:51:29 -05:00
qazal
199a36d079 add pagination to viz [pr] (#8794)
* add pagination to viz [pr]

* work

* lint
2025-01-29 04:21:53 +02:00
qazal
ba17786068 do not construct unmasked VALID (#8759)
* new lines that exist in codegen/ops

* update tests

* update sops.gz (13071 -> 13070 asts)

* fix viz too

* remove that TODO

* diff pruning

* mask assert + device

* work

* diff pruning

* re: fix viz too

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
2025-01-28 20:51:21 +02:00
qazal
3417bc1814 fix ShapeTracker spec for const [pr] (#8791) 2025-01-28 19:53:36 +02:00
nimlgen
801ec9e697 am: no hardcoded clocks (#8788)
* am: no hardcoded clocks

* better
2025-01-28 20:18:46 +03:00
qazal
e724af74d7 allow VIEW source in DEFINE_VAR spec [pr] (#8790) 2025-01-28 17:42:14 +02:00
b1tg
da464d039f fix windows ci cache (#8787)
Co-authored-by: b1tg <b1tg@users.noreply.github.com>
2025-01-28 13:22:15 +02:00
Francis Lata
dcd1941b94 fix validation test case 2025-01-28 03:06:43 -08:00
Francis Lata
fc957e7377 update validation dataloader and more cleanups 2025-01-28 03:02:17 -08:00
qazal
e8be8a5835 support lowering CONST(VIEW) in lowerer (#8785) 2025-01-28 12:04:41 +02:00
George Hotz
80089536e5 Revert "move llvm_bf16_cast to renderer for CLANG and LLVM [pr] (#8720)" (#8786)
This reverts commit af0452f116.
2025-01-28 18:59:02 +09:00
b1tg
5d62aa28dc Support CLANG backend on Windows (#8768)
* Support CLANG on Windows

* Put both backends in a windows ci

* remove coff loader

* use memmove

---------

Co-authored-by: b1tg <b1tg@users.noreply.github.com>
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
2025-01-28 18:19:34 +09:00
mesozoic-egg
af0452f116 move llvm_bf16_cast to renderer for CLANG and LLVM [pr] (#8720)
* handle bf16 via bitcasting for CLANG and LLVM

* On LLVM, skip float16 cast

* float32 on llvm lite, float32 elsewhere

* code format

* trigger pr

* move to rewriter

---------

Co-authored-by: Mesozoic Egg <mesozoic.egg@proton.mail>
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
2025-01-28 18:16:43 +09:00
nimlgen
d66680b17e hotfix: am: fix hang (#8783) 2025-01-28 11:54:19 +03:00
qazal
a65d2917cb remove unused fields from viz uop_to_json [pr] (#8782) 2025-01-28 10:50:11 +02:00
qazal
aefbc2637f test fixups from unmasked valid deletion [pr] (#8776) 2025-01-28 09:23:30 +02:00
qazal
ed672881b0 remove additions/deletion in pr + check uops are equal [pr] (#8779)
* use warnings there [pr]

* remove those + move assert_diff [pr]

* warn after log

* remove

* back
2025-01-28 08:57:34 +02:00
Ignacio Sica
2c71c60719 opt arg is int or tuple (#8780) 2025-01-28 11:02:32 +09:00
George Hotz
62655e4999 move multi into engine [pr] (#8778)
* move multi into engine [pr]

* all runtime is one sz
2025-01-28 09:15:29 +09:00
nimlgen
299fa8f37b am: unset high clocks for sleep (#8775) 2025-01-28 01:15:56 +03:00
chenyu
c99ae81f63 update default resnet LOSS_SCALER to 256 [pr] (#8774) 2025-01-27 16:59:05 -05:00
nimlgen
1c608ae34f am_smi: better spacing (#8773)
* am_smi: better spacing

* not used
2025-01-27 23:01:02 +03:00
Ignacio Sica
b240f12593 [TIP-9] rename Opt's amt to arg 2 (#8770)
* rename Opt amt to arg

* ignore_beam_cache for test_tiny

* move ignore_beam_cache to test_tiny

* move to separate pr

* revert space change

---------

Co-authored-by: chenyu <chenyu@fastmail.com>
2025-01-27 14:19:04 -05:00
chenyu
9760688e7f use IGNORE_BEAM_CACHE in search [pr] (#8772) 2025-01-27 13:41:01 -05:00
Ignacio Sica
ed1b573868 ignore beam cache in test_tiny for stateless beam (#8771) 2025-01-27 12:56:30 -05:00
Francis Lata
857e9c4c50 assert boxes again 2025-01-27 08:46:41 -08:00
Francis Lata
3b9e5a3ed4 debug test 2025-01-27 08:35:36 -08:00
Francis Lata
2177053076 Merge branch 'master' into retinanet_mlperf 2025-01-27 08:07:19 -08:00
Francis Lata
df236925c7 add debug logging for dataloader test 2025-01-27 07:57:33 -08:00
Francis Lata
e91733baae refactor on training loop and start the work on val looop 2025-01-27 07:27:19 -08:00
George Hotz
3ed146a5ff Revert "rename Opt amt to arg (#8767)" (#8769)
This reverts commit bf041659a5.
2025-01-27 23:46:37 +09:00
Ignacio Sica
bf041659a5 rename Opt amt to arg (#8767) 2025-01-27 23:36:47 +09:00
George Hotz
96bff0b4f7 contiguous is no longer needed in SGD [pr] (#8760)
* contiguous is no longer needed in SGD [pr]

* add allow condition
2025-01-27 15:19:11 +09:00
b1tg
efc7971090 add windows test to ci (#8761)
Co-authored-by: b1tg <b1tg@users.noreply.github.com>
2025-01-27 14:53:21 +09:00
George Hotz
a9d9f98d05 hotfix: those tests fail locally on mac due to buffer count 2025-01-27 07:53:48 +09:00
George Hotz
2454bf01c3 hotfix: remove shapetracker spam in viz 2025-01-27 07:20:21 +09:00
qazal
d488bbb1ec share merge_views/valid creation for CONST/DEFINE_VAR (#8758)
* share valid creation behavior for CONST/DEFINE_VAR

* work
2025-01-26 17:41:54 +02:00
qazal
bbb2dd8141 move VALID creation after merging the views (#8757)
* do valid creation later

* work for view_left

* only view(const) makes valids in view_left

* cleaner bind diff
2025-01-26 16:58:05 +02:00
George Hotz
a6e496b195 remove Function class [pr] (#8753)
* remove Function class [pr]

* actually remove function

* fix docs
2025-01-26 18:58:02 +09:00
qazal
ac70f63d4b tensor_map cleanups [pr] (#8754)
* tensor_map cleanups [pr]

* update test_schedule too
2025-01-26 11:41:54 +02:00