sirhcm/tinygrad
1
0
Fork 0
forked from tinygrad/tinygrad
Code Pull Requests 2 Actions 8 Activity
14,176 Commits 336 Branches 15 Tags
http_threading
Commit Graph
1310 Commits
Author SHA1 Message Date
qazalandGitHub 9f2a578e26 unskip TestCall.test_call_gemm_uop [pr] (#15786) 2026-04-17 16:18:51 +03:00
George HotzandGitHub 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
George HotzandGitHub a9b6cfece0 refactor llm into files (#15780) 
* refactor llm into files

* chat.html

* tokenizer cleanup

* cleanup

* tests
 2026-04-17 12:33:11 +08:00
George HotzandGitHub ec00cefa5b llm is the only app (#15779) 
* tinygrad/llm is the only app

* upd pyproject

* claude refs

* scoping

* min diff
 2026-04-17 10:44:48 +08:00
chenyuandGitHub f0c12a2004 another form of assign to itself (#15770) 2026-04-16 15:17:19 -04:00
chenyuandGitHub d147e2a549 update test_nested_after_contiguous_store (#15763) 
add kernel counts and some TODOs
 2026-04-16 09:59:26 -04:00
George HotzandGitHub f57380cbc2 simplify GatedDeltaNetBlock using two state tensors (#15704) 
* test double after

* simpler ssm

* no double test
 2026-04-16 21:14:00 +08:00
George HotzandGitHub d1cce7a476 put the ranges on store instead of after (#15759) 
* put the ranges on store instead of after

* better assert

* fix stuff

* comment out slow rules i don't understand

* simpler rule

* closer

* return false for store

* fix loop

* only a few schedule failures remain

* remove stores to self

* all tests pass locally

* remove junk

* regression test and fix

* better test, bump broken torch count

* bugfix with regression test

* new fusion is better
 2026-04-16 19:06:40 +08:00
George HotzandGitHub d24466c844 CALL with return value is FUNCTION (#15758) 
* CALL with return value is FUNCTION (GPT try)

* cleanups
 2026-04-16 13:25:07 +08:00
chenyuandGitHub 10c262ced8 update tests that use UOp.size (#15753) 2026-04-15 21:58:27 -04:00
George HotzandGitHub 1ae6528bb6 move schedule into schedule (#15736) 
* move schedule into schedule

* callify to root

* sched docs
 2026-04-15 11:03:25 +08:00
wozeparrotandGitHub 2b8d303f75 allreduce in precast dtype (#15689) 2026-04-13 20:24:12 -07:00
George HotzandGitHub 7610bdc59e block multistore, it's not supported (#15708) 2026-04-13 20:57:59 +08:00
George HotzandGitHub 4c1fb18a09 Revert "Revert "Tests for GatedDeltaNetBlock + fix multi after assign issue (…" (#15703) 
This reverts commit 0cec42db71.
 2026-04-13 19:09:38 +08:00
George HotzandGitHub 0cec42db71 Revert "Tests for GatedDeltaNetBlock + fix multi after assign issue (#15700)" (#15702) 
This reverts commit 6f5d756282.
 2026-04-13 19:06:44 +08:00
George HotzandGitHub 6f5d756282 Tests for GatedDeltaNetBlock + fix multi after assign issue (#15700) 
* broken after/assign test

* test for GatedDeltaNet

* better comments

* fix issue 1 with multi kernel

* fix 2

* fix

* linter

* public api + cleanup
 2026-04-13 18:43:23 +08:00
chenyuandGitHub e706f408cb suppress test warnings from numpy (#15688) 2026-04-11 22:33:20 -04:00
Graham RobbinsGitHubGeorge Hotz
 		4ca844e96b add Q1_0 gguf type (#15683) 
* add Q1_0

* better description

* fix trailing whitespace

---------

Co-authored-by: George Hotz <[email protected]>
 2026-04-11 18:17:24 +08:00
wozeparrotandGitHub 457508d5a0 llama: save more 2 (#15681) 2026-04-11 01:03:36 -07:00
George HotzandGitHub b5a9465b13 llm: add support for moonlight (deepseek MLA) (#15466) 
* add gguf Q5_0

* it works

* rebase

* simpler test

* class

* less diff

* dicts

* normal names

* simplify

* this

* simpler

* work

* work
 2026-04-11 10:32:48 +08:00
George HotzandGitHub 9092f2a8c0 llm: add shared_expert and rope_dim support from qwen35 (#15673) 
* llm: add shared_expert and rope_dim support from qwen35

* refactor into FFNBlock and TransformerBlock

* norms where they belong
 2026-04-10 19:18:27 +08:00
George HotzandGitHub 35e3983840 Add Q5_0, Q5_1, and bfloat16 GGUF types (#15644) 2026-04-08 17:16:19 +08:00
sirhcmandGitHub acf239e4d2 specify renderer in DEV, <dev>_<ren>=1 is deprecated (#15551) 2026-03-31 18:35:14 -04:00
nimlgenandGitHub 5181c8e23a llm: fix nan in kvcache (#15552) 2026-04-01 00:38:45 +03:00
b1tgandGitHub a63392a565 llm: pairwise ranking topk for MoE expert selection (#15499) 2026-03-31 12:46:39 +08:00
nimlgenandGitHub 0d6fc0f571 jit: graphing in uops (#15489) 
* jit: graphing as rewrite rule

* f

* +metal,cuda

* x

* cl

* x

* x

* simpler

* f

* m

* x

* revert?

* revert2

* back

* back

* t

* x

* m

* x

* c

* x

* l

* x

* comment

* smaller

* rv

* x

* x
 2026-03-27 19:09:02 +03:00
sirhcmandGitHub bc180a963c deprecate <dev>=1 in favor of DEV=<dev> (#15467) 
* start work on target

* add test

* update actions to use DEV

* update docs

* update readmes

* tests need that too

* update example

* update tests (comments)

* fix that test

* ruff

* mypy

* oops

* remove getenvs

* don't add Target yet

* and the test

* lint

* and docs

* more stuff

* assert

* few more fixes

* test assert
 2026-03-26 03:48:03 -04:00
George HotzandGitHub fe2690399b llm: support assistant prefill + refactor to TransformerConfig (#15457) 
* llm: support assistant prefill

* refactor to ModelConfig

* TransformerConfig

* more
 2026-03-25 10:50:48 +08:00
George HotzandGitHub a33ac869aa llm server: temperature + test client (#15444) 
* improvements to the llm server

* eval script

* eval llm

* better eval gets 58.71

* cleanups

* add temperature, but multinomial is absurdly slow

* claude is so smart

* lint

* remove slop

* no more stop
 2026-03-24 21:07:15 +08:00
chenyuandGitHub c491345766 pass device into Tensor._frompy (#15385) 
* pass device into Tensor._frompy

with this, canonicalize_device is the only usage of Device in tensor.py

* export_model.py
 2026-03-20 05:09:01 -04:00
George HotzandGitHub 3b75d8a7a2 fix double after bug in rangeify (#15381) 2026-03-20 14:53:46 +08:00
sirhcmandGitHub 0c89340a1e automatically emulate unsupported (tiny) floats [skip_process_replay] (#15366) 2026-03-20 02:31:44 -04:00
chenyuandGitHub bf33c5f796 remove gradient materialize_grads (#15367) 
effectively default to True

and removed *0 hack in Tensor.copysign. now dy/dx=0 if y does not depend on x

remove
 2026-03-19 23:36:03 -04:00
chenyuandGitHub fceb21c315 Tensor(uop) uses device from uop (#15340) 2026-03-18 02:56:06 -04:00
George HotzandGitHub 9d95321be3 set allow_implicit=False by default (#15319) 
* set allow_implicit=False by default

* modernize beautiful mnist
 2026-03-17 17:14:38 +08:00
George HotzandGitHub 584ec75aa2 precompile backward (#15311) 
* add precompile backward support

* cleanups

* fix

* compact grad

* split v not split

* simpler

* no NOOPT
 2026-03-17 15:28:40 +08:00
b1tgandGitHub 856a839efc llm: fix qwen3 moe topk renormalization (#15201) 2026-03-17 12:57:33 +08:00
George HotzandGitHub 3ff03be413 call always has tuple (#15297) 
* call always has tuple

* fix pre-commit and simplify

* update

* fix

* move that assert

* tuple

* fix multi

* cleanups

* fix merge
 2026-03-17 10:58:46 +08:00
wozeparrotandGitHub 674c760974 embedded bwd vocab shard (#15001) 
* fix: remove more multi from call

* feat: embedding bwd vocab sharding

* clean: unused import

* clean: don't actually need this pattern
 2026-03-16 19:37:16 -07:00
chenyuandGitHub 02afb45f29 remove UOp.assign [pr] (#15300) 
* remove UOp.assign [pr]

it's all store and after, UOp is immutable

* fix test
 2026-03-16 21:45:41 -04:00
chenyuandGitHub 3e2b7803e6 view assign replaces at buffer identity (#15298) 
matches what functions capture
 2026-03-16 19:58:38 -04:00
George HotzandGitHub 476276f4b4 support grads on tuples (#15287) 
* support grads on tuples

* simpler

* grad_fxn works

* cleanups

* unused
 2026-03-16 17:39:34 +08:00
George HotzandGitHub 08662bc4ab add TUPLE/GETTUPLE, simple tests pass (#15286) 
* simple tuple stuff passes

* resolved
 2026-03-16 15:06:02 +08:00
chenyuandGitHub cd14e8e64b allocations contiguous is store+after (#15280) 2026-03-15 11:58:40 -04:00
Sieds LyklesandGitHub 4b59083d7c assign into empty works (#15256) 2026-03-13 10:24:29 -04:00
chenyuandGitHub 018c01508d test case for call precompile multi (#15254) 2026-03-13 06:28:43 -04:00
b1tgGitHubb1tgchenyu
 		18dc77ccab add fp8 fnuz dtypes with PYTHON backend support (#14945) 
* add fp8 fnuz dtypes with PYTHON backend support

* rm emu related change

* clarify fp8 fnuz zero handling

* Revert "rm emu related change"

This reverts commit efa4763c22.

---------

Co-authored-by: b1tg <[email protected]>
Co-authored-by: chenyu <[email protected]>
 2026-03-11 22:30:18 -04:00
George HotzandGitHub 4f3f55328b do not patch on invalid tensor tests (#15226) 
* do not patch on invalid tensor tests

* cleanup
 2026-03-12 09:35:20 +08:00
sirhcmandGitHub 2fb8a7f60f fix test_invalid_tensor when before values are nan (#15215) 2026-03-10 23:51:19 -04:00
sirhcmandGitHub ffaafd391a Invalid in Tensor (#15154) 2026-03-10 02:49:54 -04:00