nimlgen/tinygrad
1
0
Fork 0
forked from tinygrad/tinygrad
Code Pull Requests Activity
13,471 Commits 334 Branches 15 Tags
python_prg
Commit Graph
13471 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
George HotzandGitHub 41b2a25be6 Fix exponential behavior in lazyops (#2890) 
* add cache to ast_parse and lazyop builder

* add caches
 2023-12-20 22:06:50 -08:00
George HotzandGitHub 8c4a0f8e15 Fix int child count (#2882) 
* pad ops broke coder

* that contiguous fixes it

* Update lazy.py

* recursive add

* fix all

* revert that

* todo test
 2023-12-20 21:06:27 -08:00
chenyuandGitHub 8a04107d30 move the op casting logic from mlops to tensor try 2 (#2887) 
* unary works

* where works

* add sub mul

* xor div

* CMPLT

* sparse_categorical_crossentropy

* image const

* sparse_categorical_crossentropy
 2023-12-20 23:50:37 -05:00
George HotzandGitHub 7da2325dc7 get_lazyops() -> lazyops (#2884) 
* get_lazyops() -> lazyops

* don't compare empty mem
 2023-12-20 18:04:49 -08:00
George HotzandGitHub 64dded27f0 pad ops broke coder (#2881) 
* pad ops broke coder

* that contiguous fixes it

* Update lazy.py
 2023-12-20 17:03:41 -08:00
George HotzandGitHub e1861ab65e remove realize from optimizer (#2880) 
* remove realize from optimizer

* one still needed

* opt realize
 2023-12-20 16:42:41 -08:00
George HotzandGitHub 1765849937 new lazy, benchmark (#2878) 
* lazy rewrite, try 2

* min fix tests

* pass contig test

* put broken pads back

* move that to realize

* no contig child fixes array packing

* so wrong

* now that's correct

* base children

* fix bind issues

* disable to_image_idx

* fix tests

* that failure shouldn't break other tests

* more fixes

* fix torch

* skip failing tests in CI

* 1e-7

* half is broken

* 1e-6 margin of error
 2023-12-20 14:33:21 -08:00
Peter CawleyandGitHub dae8976889 Fix reshape merging with masks (#2877) 2023-12-20 14:00:58 -08:00
George HotzandGitHub 8fe24038d8 Revert "mulacc fusion cleanup (#2871)" (#2876) 
This reverts commit 863c5b26ed.
 2023-12-20 13:26:25 -08:00
qazalandGitHub 863c5b26ed mulacc fusion cleanup (#2871) 
* add mulacc fusion tests

* cleanup the implementation

* fix indent in the test utility

* less verbose
 2023-12-20 15:39:54 -05:00
chenyuandGitHub e13b4964d7 remove the all_int(shape) check in Tensor._loadop (#2874) 
* remove the all_int(shape) check in Tensor._loadop

we can support jittable symbolic shape random with custom rand now, and we can formalize it in the test after threefry is ready

* MOCKHIP false positive
 2023-12-20 15:04:50 -05:00
qazalandGitHub 5f07ef455e update dtypes (#2872) 2023-12-20 15:04:02 -05:00
chenyuandGitHub 857c35d256 make gpt2 decode output just once at the end (#2869) 
also updated function name from greedy_until to generate, as it's not greedy nor until
 2023-12-20 12:14:55 -05:00
chenyuandGitHub e92069fb1c remove unused symbolic.is_sym_int (#2868) 2023-12-20 11:37:54 -05:00
George HotzandGitHub ca59054463 fix shapetracker math (#2861) 
* proper test

* all st math good now

* fix real_strides bug
 2023-12-19 22:17:34 -08:00
chenyuandGitHub 5a739e8c20 update one skipped pad_reshape test that was fine (#2860) 
* update one skipped pad_reshape test that was fine

had a typo

* this one passed
 2023-12-19 23:25:52 -05:00
chenyuandGitHub 39af93ed7c minor tensor.py function cleanup (#2859) 
* minor tensor.py function cleanup

* where outputs not aligned yet
 2023-12-19 22:39:39 -05:00
geohot 94f71fe238 random and empty shouldn't reshape 2023-12-19 18:09:03 -08:00
geohot 637879af78 add direct install to readme 2023-12-19 18:04:00 -08:00
chenyuandGitHub ad233d557f disable reshape merging with masks (#2858) 
fuzzer found a bug, and it's not complete
 2023-12-19 19:06:16 -05:00
chenyuandGitHub 1231ec5a02 run the sz.py line count at the end of linter ci (#2857) 2023-12-19 16:33:12 -05:00
geohot ac6ec936cd update contributing 2023-12-19 12:19:14 -08:00
geohot e477cc2f45 hotfix: README is ~25 ops to stop getting PRs about it 2023-12-19 11:53:35 -08:00
Oleg RybalkoandGitHub 42a038c83f More readable torch_load ext check (#2853) 
* more readable extension check

* enable tarfile test

* detach tensor if requires grad in torch
 2023-12-19 14:53:15 -05:00
chenyuandGitHub 172a88e719 skip slow test_indexing on METAL (#2852) 
LLVM still runs and is a lot faster, would be curious to know why.
also reworded some error messages and remove regex check
 2023-12-19 12:00:54 -05:00
chenyuandGitHub 6d7e9e0a56 hotfix convert Y_train to int before passing into index (#2850) 2023-12-19 11:40:56 -05:00
geohotstanandGitHub fec8e9060c Add simple fancy indexing exceptions (#2706) 
* fancy indexing raise error

* updated error message

* improved error check

* oops

* fixed onnx

* oops typo

* merge

* add full_flatten

* try

* merged and updated some tests

* more cleaning

* done

* temp fix onnx

* try

* add todo in onnx_test

* reword

* gah
 2023-12-19 11:23:51 -05:00
qazalandGitHub 417d42a363 UOps support for ImageDType (#2848) 
* cleanup buffer dtypes in global_load

* update with feedback
 2023-12-19 09:39:48 -05:00
geohot 90fb09b55c remove unused _device_extra_args 2023-12-18 22:14:58 -08:00
George HotzandGitHub b2192b5400 minor improvements (#2845) 2023-12-18 22:09:08 -08:00
geohot d086325b1b hotfix: failing tests 2023-12-18 21:12:42 -08:00
George HotzGitHubqazalqazal
 		07df14aa0e HIP cleanups (#2843) 
* move everything to code_for_op to reason about it

* loop the loopable parts

* its not that unreadable

* these are loopable too

* nitpick

* tests p1 - replace these with the actual compiler running alu ops tests

* tests p2: compile test_dtype_alu in HIP!

+add to CI

* nobody liked test_renderer

* revert test_dtypes change

* isolated mockhip tests

* dont need the WHERE hack after #2782

+ruff

* bf16 is broken in HIP

job failed in: https://github.com/tinygrad/tinygrad/actions/runs/7232101987/job/19705951290?pr=2778#step:8:73

* picking this back up

* add compile tests for unary ops and binary ops

* MOD is only in ints

* CMPLT wont work after the dtypes pr is merged because it will always be bool

* test all combinations

* Update cstyle.py

* don't use vload

* no getenv

* set seed

---------

Co-authored-by: qazal <[email protected]>
Co-authored-by: qazal <[email protected]>
 2023-12-18 21:09:32 -08:00
geohot b6d71b131e hotfix: push broken tests 2023-12-18 21:08:42 -08:00
geohot 9b35186bbe hotfix: don't import from tinygrad in sz.py 2023-12-18 20:49:46 -08:00
George HotzandGitHub 6617dcf095 move graph to runtime, check line count with sz.py (#2842) 
* move graph to runtime, check line count with sz.py

* oops, didn't save

* dtype aliases

* restore comment, REALCOUNT
 2023-12-18 20:30:06 -08:00
chenyuandGitHub 15dc5bcfbd remove cstyle hip functions that deals with different types of input (#2840) 2023-12-18 22:43:22 -05:00
chenyuandGitHub dad9253d52 minor clean up in kernels (#2832) 
clean up some long lines and combined some short lines
 2023-12-18 19:35:59 -05:00
geohot 954a2fef75 hotfix: add JITGRAPH and invert sints 2023-12-18 16:33:22 -08:00
George HotzandGitHub 80f53245e8 shapetracker add and invert (#2828) 
* invert (broken)

* decent invert

* shapetracker invert works

* plus is meh, invert is good

* support invert mask

* a few more invert tests

* shapetracker math invert test
 2023-12-18 16:03:27 -08:00
chenyuandGitHub 73cadfbb3c Remove pytest markers (#2831) 
* remove pytest marker

* fix some, skip some

* tweak

* fix

* skip slow

* skip more
 2023-12-18 18:53:28 -05:00
chenyuandGitHub 264fe9c93f clean up test_dtype.py (#2827) 
make is_dtype_supported a pure function and clean up long lines
 2023-12-18 16:06:09 -05:00
chenyuandGitHub 20ea43b6e7 dtypes.from_py to convert py types to dtypes (#2826) 
also updated some tests to test against default dtypes
 2023-12-18 14:23:31 -05:00
chenyuandGitHub 0723f26c80 dtypes.default_float and dtypes.default_int (#2824) 2023-12-18 12:21:44 -05:00
chenyuandGitHub 8aab19ce3d Tensor.full of bool has dtypes.bool (#2823) 2023-12-18 10:51:17 -05:00
chenyuandGitHub 220abcd8ff fix squeeze of 0-dim Tensor with negative dim (#2821) 
if ndim=0, only accepted dim is 0, -1, None. other negative dim results in IndexError
 2023-12-17 22:02:07 -05:00
chenyuandGitHub 21ec7e09f6 minor cleanup of image (#2820) 
use transpose for transpose instead of permute, and use pad for pad instead of slice
 2023-12-17 20:53:49 -05:00
chenyuandGitHub 959d9cfed4 clean up ops_torch and ops_cpu (#2819) 2023-12-17 19:35:19 -05:00
Rory ClearGitHubchenyu
 		f409b57854 update metal matmul and matvec for new device style (#2732) 
* update for new device style

* create device before compile

---------

Co-authored-by: chenyu <[email protected]>
 2023-12-17 16:15:07 -05:00
chenyuandGitHub 91adb119b8 remove match_type in ops_torch and ops_cpu (#2817) 
* remove match_type in ops_torch and ops_cpu

input dtypes are aligned and casted in mlops

* dict union only after python3.9

* fix that

* fix Sigmoid forward cast
 2023-12-17 15:32:30 -05:00
Maksym SobolyevandGitHub 887f3d9933 Make torch backend more usable, fix bfloat support in the llvm backend (#2765) 
* Uncripple dtype tests, TestBFloat16DType never actually runs.

* Fix conversion from/to bfloat16.

Call cast() recursively, so that it works for any type combo.

* Run this test on torch backend as well.

* Add torch.bfloat16.

* Add support for ushort and uint.

* Convert np.uint32 to np.int32 when loading.

* Fix warning.
 2023-12-17 14:04:26 -05:00