nimlgen/tinygrad
1
0
Fork 0
forked from tinygrad/tinygrad
Code Pull Requests Activity
13,974 Commits 334 Branches 15 Tags
2864036e8ef99cab0a14d3bc19574a052d09b1b8
Commit Graph
409 Commits
Author SHA1 Message Date
sirhcmandGitHub f3a5337825 correct spelling of coalesce (#17103) 2026-07-20 21:44:04 -04:00
chenyuandGitHub 13ca9bd8a6 remove dtypes.index again (#17104) 
also reverted some dtype change, the split made things needlessly complicated
 2026-07-20 20:30:04 -04:00
sirhcmandGitHub 980748ccfc add multiple_of to ParamArg (#17101) 2026-07-20 20:11:54 -04:00
George HotzandGitHub f7ce7f330d llm: minor fixes + tests (#17099) 
* llm: minor fixes + tests

* error
 2026-07-20 14:31:20 -07:00
Pol Puigdemont PlanaandGitHub ef77963cfd derivative of logsumexp is independent of max (#17088) 
same as #7009 but for logsumexp and logcumsumexp.
fwd+bwd kernel count 5 -> 3 for both. gradients unchanged
(ties, -inf masks, torch-compared at grad_atol=1e-7).
 2026-07-20 06:52:16 -07:00
chenyuandGitHub 0146a30125 improve cast to unsign min_max [pr] (#17078) 2026-07-18 21:58:41 -04:00
George HotzandGitHub b53cd35cff llm: make tokenizer fast (kimi) (#17077) 
* llm: make tokenizer fast

* simpler

* re.escape + qcom mypy fix
 2026-07-18 17:31:59 -07:00
chenyuandGitHub f315df29a0 no weak Tensor from and to real buffer (#17067) 
* no weak Tensor from and to real buffer

creation, assign, safe_save

* is_numpy_ndarray to tensor

* one more
 2026-07-17 16:09:10 -04:00
George HotzandGitHub 86a6ad8ed2 llm: split cli.py into serve.py with the HTTP server (#17065) 
* llm: split cli.py into serve.py with the HTTP server

* min edit
 2026-07-17 10:45:37 -07:00
George HotzandGitHub 3ee2baf71d llm: add tool calling support (kimi) (#17061) 
* llm: add tool calling support

* simpler

* cls

* gpt cleanup

* more gpt cleanups

* tests for tools calling
 2026-07-17 10:20:01 -07:00
George HotzandGitHub 46172bb7c7 llm: add optional jinja template support (kimi) (#17058) 
* add jinja template support (kimi)

* fix tests

* lil

* more crap to fallback
 2026-07-16 19:02:58 -07:00
George HotzandGitHub 61e104bdfb use UOp.wmma everywhere (#17045) 
* use UOp.wmma everywhere

* fix
 2026-07-16 10:40:48 -07:00
chenyuandGitHub 5a4156c5d1 bitcast and element_size raise for weak dtypes (#17046) 2026-07-16 13:07:45 -04:00
chenyuandGitHub be075b200a weak dtypes in dtype_from_uop [PR] (#17032) 
* weak dtypes in dtype_from_uop [PR]

* no weak in spec_program

* weak const fold tests
 2026-07-15 16:54:31 -04:00
nimlgenandGitHub d6fddb066f usb: keep only custom (#17029) 
* usb: keep only custom

* mockgpu by gpt

* gpt said sorry

* revert

* reset

* fix

* flash
 2026-07-15 22:40:52 +03:00
chenyuandGitHub 0d19970edc least_upper_dtype in dtype_from_uop [PR] (#17028) 2026-07-15 09:27:53 -04:00
chenyuandGitHub 47ddf94f17 remove InvalidType lt and gt (#17023) 
not really used
 2026-07-14 21:59:15 -04:00
sirhcmandGitHub c9baa2ef79 use pattern matcher in contiguous_view_offset [PR] (#17022) 2026-07-14 19:37:36 -04:00
chenyuandGitHub 0410c9325d make test/null follow the SPEC (#17010) 2026-07-13 14:01:41 -04:00
George HotzandGitHub 03ecad9486 full removal of dtype.vec (#16996) 
* full removal of dtype.vec

* fix typo
 2026-07-12 09:25:50 -07:00
chenyuandGitHub 5a0751076d simpler vconst_like (#16990) 
also always use const_like in symbolic, fixed a crash
 2026-07-12 08:21:57 -04:00
chenyuandGitHub 047a467bf9 delete UOp._stack and UOp.vectorize [PR] (#16988) 2026-07-11 15:26:55 -04:00
chenyuandGitHub 9d47014fd8 first class STACK [PR] (#16986) 2026-07-11 13:55:32 -04:00
chenyuandGitHub 2fda6b3888 fix shape broadcast for symbolic (#16970) 
* fix shape broadcast for symbolic

0 or smax was wrong for the unresolved case

* test with null
 2026-07-10 15:13:01 -04:00
chenyuandGitHub 8d546f55e9 dtypes.weakfloat (#16956) 2026-07-09 19:32:21 -04:00
George HotzandGitHub d80c971ea2 param + cast dtype in arg (#16955) 
* param + cast dtype in arg

* require dtype

* fix replace in dtype decomps

* fix x86

* dtype none is fine there

* Revert "dtype none is fine there"

This reverts commit 5315cc3e59.
 2026-07-09 16:32:01 -07:00
chenyuandGitHub 78223d690a update SPECIAL spec (#16952) 
also updated tests with bad UOp
 2026-07-09 19:05:34 -04:00
George HotzandGitHub 95681f17ee remove extra dtypes from tests (#16954) 
* remove extra dtypes from tests

* a few more
 2026-07-09 14:59:56 -07:00
George HotzandGitHub 8d4c9d1058 make the upat compiler match the spec (glm) [PR] (#16948) 
* make the upat compiler match the spec

* cleanups

* CUSTOMI/CUSTOM

* fix that test

* pyliteral

* fix variable shadowing
 2026-07-09 12:30:39 -07:00
chenyuandGitHub ba2c68b1ed start dtypes.index [PR] (#16949) 2026-07-09 15:12:23 -04:00
sirhcmandGitHub da435c719b tinyfs: use assign (#16947) 2026-07-09 14:52:42 -04:00
George HotzandGitHub 904b51a783 dtype fixups (grok) [PR] (#16940) 
* dtype fixups

* a few more

* amd fixes

* revert that

* Remove comment in test_graph_rewrite_div_folding_bug

Removed comment about STACK dtype matching its element dtype in the test for graph rewrite division folding bug.
 2026-07-09 11:03:22 -07:00
George HotzandGitHub fdffc6c0c8 remove dtypes base (#16931) 
* remove dtypes base

* find/replace bug

* vcount is junk too
 2026-07-08 11:48:11 -07:00
George HotzandGitHub 63cb1369cb remove ImageDType (glm) (#16918) 
* remove ImageDType (glm)

* gpt work

* better

* is_image_shape

* cleanups

* remove crap from index

* revert crap python change

* image index with float, this needs to be fixed

* simp
 2026-07-08 11:27:52 -07:00
George HotzandGitHub 6e05cbdbcb remove the dtype.scalar calls everywhere [pr] (#16930) 
* remove the dtype.scalar calls everywhere

* remove base everywhere

* Revert "remove base everywhere"

This reverts commit 0dd0a5a243.
 2026-07-08 10:10:03 -07:00
chenyuandGitHub c26468c5d0 keccak and hash to mixin [PR] (#16925) 2026-07-07 21:12:46 -04:00
chenyuandGitHub c43a3fdebb bitcast to mixin [PR] (#16924) 2026-07-07 20:03:29 -04:00
George HotzandGitHub d8fbbff260 EXPAND adds dims to the front (glm) (#16908) 
* EXPAND adds dims to the front (glm)

* more read images

* simpler + spec

* spec should use n instead of s'

* bump comma to 12
 2026-07-07 11:03:06 -07:00
George HotzandGitHub 0035bb6fa8 only front reduce (#16901) 
* only front reduce

* lint + spec

* that should fail

* meh
 2026-07-07 09:55:07 -07:00
qazalandGitHub f0a06fae4c viz: readable REWRITE_ERROR (#16903) 
* viz: readable REWRITE_ERROR

* traceback.format_exc handles this
 2026-07-07 17:25:42 +09:00
qazalandGitHub 9454e2ddbf viz: inf loop early cutoff (#16902) 
* pause on error

* reorder err

* simpler
 2026-07-07 17:00:45 +09:00
George HotzandGitHub 40d112d4d6 remove ptr=True from index (#16898) 2026-07-06 18:34:18 -07:00
George HotzandGitHub 6b8b2f5aeb remove PtrDtype from tests (#16896) 
* remove PtrDtype from tests

* fixes

* not needed

* no -1
 2026-07-06 18:22:22 -07:00
George HotzandGitHub 9512dc30f4 remove vec from const (#16889) 
* remove vec from const

* reject in spec

* remove all vector dtypes

* not needed

* remove that

* clean up invalid
 2026-07-06 12:23:07 -07:00
George HotzandGitHub c92a4c0442 move devectorizer remains into coalese (#16869) 
* move devectorizer remains into coalese

* fix flaky
 2026-07-04 15:36:45 -07:00
George HotzandGitHub a145fdce4b remove the .gep method (#16867) 2026-07-04 14:07:09 -07:00
George HotzandGitHub 3fd6f3d28c no more GEP (#16866) 2026-07-04 13:42:28 -07:00
George HotzandGitHub d76bcbaa02 deletions after new codegen (#16864) 
* deletions after new codegen

* delete the expander
 2026-07-04 13:12:11 -07:00
George HotzandGitHub c7e7687bd3 new codegen, try 3 (#16781) 
* new codegen, try 3 [pr]

* reduce

* diffs from cg2

* minor fixes

* mergable

* new local buffers

* fixes

* correct fix

* add barrier, and image is after that

* fix get

* wmma work

* devec wmma

* fixes for wmma

* slightly more flexible

* fix tensor cores

* fix linter

* tests

* remove invalid?

* remove extra

* wmma merge

* flip reshape/permute on wmma

* broadcast_binary None

* update symbolic for vecless

* x86 fix

* fix custom llama kernel

* is_ptr hack

* fix pre-commit

* fix wmma

* simpler

* fixes

* better hreduce

* move reduce axes

* all wmma tests pass

* allow stacked wmma

* permute cleanups

* late permute

* correct permute flip

* fix group for reduce

* fix wmma permuted

* all pass

* mop cleanup

* new test

* push permute/reshape

* revert that

* half
 2026-07-04 12:55:44 -07:00
chenyuandGitHub 709251e22d more generic fold_add_divmod_recombine [pr] (#16854) 2026-07-03 22:00:59 -04:00