github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-08 22:48:25 -05:00
Code Issues Packages Projects Releases Wiki Activity
9,095 Commits 203 Branches 13 Tags
cfa65bea050cdec0586ef0aa73a848643a2b7e2a
Commit Graph

9095 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Sieds Lykles
cfa65bea05 Subtract 1 from Variable upper bound (#10715) 2025-06-09 09:25:53 -07:00
George Hotz
ef58ab340a hotfix: remove n=auto from REMOTE=1 test 2025-06-09 09:19:36 -07:00
qazal
419a1286f2 viz: share cacheKey [pr] (#10717) 2025-06-09 17:48:29 +03:00
chenyu
35523dc35f move BLOCK_REORDER to caller [pr] (#10711) 
so block_reorder tests won't fail with flag set to 0
 2025-06-08 23:26:01 -04:00
chenyu
bb34c28b36 debug flag for linearize block_reorder [pr] (#10710) 2025-06-08 22:26:06 -04:00
chenyu
d93a0bee6b mlperf ci uses its own cache (#10705) 
not to interfere with regular cache which is used by benchmark
 2025-06-08 19:43:32 -04:00
qazal
8cdf6e4d1e viz memory graph tiny fixes [pr] (#10709) 
* sched_sink is a step

* offset for yaxis

* clear existing

* scale offset
 2025-06-09 01:10:12 +03:00
George Hotz
81b9c04574 move high level stuff to unit tests [pr] (#10708) 
* move high level stuff to unit tests [pr]

* process replay on unit tests

* fix pr, less compute

* set omp num threads

* set 200MB buffer size limit

* delete junk

* fix tests

* faster

* move test_indexing to unit

* faster
 2025-06-08 14:05:56 -07:00
nimlgen
171580e9ec am: fix reg update (#10707) 2025-06-08 21:45:55 +03:00
George Hotz
4305f532d9 clean up apt stuff (#10706) 
* clean up apt stuff

* single apt install

* fixes

* fix opencl + ldconfig
 2025-06-08 11:06:09 -07:00
George Hotz
4e2c3560b4 smaller tests are faster tests [pr] (#10704) 
* remove del spam from CI

* more

* preconstruct default buffer spec

* ignore those errors

* check exception

* more exception check

* skip stuff

* smaller tests mean faster tests

* a few more
 2025-06-08 10:54:19 -07:00
George Hotz
67a1c92fc0 remove del spam from CI (#10699) 
* remove del spam from CI

* more

* preconstruct default buffer spec

* ignore those errors

* check exception

* more exception check

* skip stuff
 2025-06-08 10:14:30 -07:00
George Hotz
32141ec867 make apt CI faster (#10702) 2025-06-08 09:43:39 -07:00
chenyu
4f535641f7 add one huggingface_onnx test to mac benchmark ci (#10700) 
this crashed for me on onnx parser pr but seems fine for the author. see if ci mac is fine
 2025-06-08 12:26:12 -04:00
George Hotz
32e9949052 rename lazydata to uop (#10698) 2025-06-08 08:42:22 -07:00
uuuvn
8e3f337075 Skip flaky test in ci (#10696) 
`test_data_parallel_resnet_train_step` is already skipped on LLVM/CPU:

```python
@unittest.skipIf(CI and REAL_DEV in ("CUDA", "NV", "LLVM", "CPU"), "slow, and flaky on LLVM/CPU")
@unittest.skipIf(REAL_DEV == "WEBGPU" and not OSX, "WEBGPU Vulkan can only run kernels with up to 10 buffers")
def test_data_parallel_resnet_train_step(self):
```

It looks like `test_data_parallel_resnet` (no `_train_step`) is flaky in a similar way:
https://github.com/tinygrad/tinygrad/actions/runs/15472667248/job/43560773882?pr=10642#step:9:64
 2025-06-08 08:24:09 -07:00
George Hotz
3ece2e4bb5 hotfix: remove accel from extra 2025-06-08 08:20:34 -07:00
qazal
1ad8062591 more generic naming in VIZ [pr] (#10695) 
* note

* rename kernel to ctx

* rename uop things to currentStep + expandSteps

* already destructured

* some things that were called ctx are steps

* still a kernel
 2025-06-08 15:37:39 +03:00
qazal
c70486908e viz: clicking a KERNEL node can open codegen rewrite (#10683) 
* work

* now it doesn't have 20% slowdown

* label like this

* closer

* ansiStrip

* remove

* better

* id is faster

* fix that
 2025-06-08 13:11:03 +03:00
George Hotz
48eb7d76b1 use ALLOW_DEVICE_USAGE context variable instead of MainProcess check (#10693) 
* use DISALLOW_DEVICE_OPEN context variable instead of MainProcess check

* device usage can be disallowed
 2025-06-08 00:07:40 -07:00
geohotstan
dedff0e96c fix run huggingface onnx debug (#10679) 2025-06-08 00:59:20 -04:00
George Hotz
8c76250d31 speed up a few tests (#10692) 2025-06-07 20:39:25 -07:00
chenyu
e80870e27c BasicBlock2 -> BasicBlock [pr] (#10691) 2025-06-07 23:33:51 -04:00
George Hotz
7ff175c022 cache a venv to avoid pip usage (#10689) 
* try built in pip caching

* try venv

* export venv

* set VIRTUAL_ENV

* revert that

* venv key

* fix

* ci cache hit?

* fix windows
 2025-06-07 20:13:41 -07:00
ihar
40c1479267 added unit tests for 'argfix' (#10678) 2025-06-07 22:17:10 -04:00
ihar
74b849b5e1 remove unnecessary 'argfix' because 'view' is an alias to 'reshape'. all functionality must be inside 'reshape' (#10677) 
* remove unnecessary 'argfix' because 'view' is an alias to 'reshape'. all functionality must be inside 'reshape'

* added the same set of unit tests for 'view' as for 'reshape' since 'view' is just an alias for 'reshape'

* improved tests for 'view' op
 2025-06-07 22:15:31 -04:00
chenyu
e88fe41d37 update vits vctk model to use download from huggingface (#10688) 
google drive points to a warning page that does not work
 2025-06-07 20:47:28 -04:00
Sieds Lykles
c29a56dd51 Fix whisper OOB (#10685) 
* fix whisper and test

* remove import
 2025-06-07 20:23:50 -04:00
George Hotz
53ed64e133 ci speed work 1 (#10676) 
* skip a few slow tests

* use a venv for python packages

* create venv

* no user, it's in venv

* ignore venv

* venv

* new cache key

* try that

* this

* version the python cache
 2025-06-07 16:33:11 -07:00
George Hotz
db01c5a08a ramp.py file from stream (#10686) 2025-06-07 14:58:21 -07:00
Sieds Lykles
2f605eadf7 fix oob (#10666) 2025-06-07 11:32:03 -04:00
qazal
cb61774ab6 move shared viz fields out of serve.py [pr] (#10684) 
* move shared viz fields out [pr]

* update javascript

* update test_viz
 2025-06-07 17:18:18 +03:00
qazal
b515d796fb inline viz get_name [pr] (#10682) 
* inline viz get_name [pr]

* changing name_fxn makes this simpler

* waitUntil dom
 2025-06-07 11:16:16 +03:00
qazal
86a19e19e8 cleanup bits of viz [pr] (#10681) 2025-06-07 09:18:12 +03:00
wozeparrot
e3805171e2 feat: variable bs bitcast (#10674) 2025-06-06 17:21:53 -07:00
George Hotz
54db1f8ee8 prevent huge waste of multi ram (#10669) 
* prevent huge waste of multi ram

* fix ram usage

* only define var

* add resolve

* fix tests

* fix cifar training

* remove that logic

* fix test without long
 2025-06-06 17:17:21 -07:00
George Hotz
b68b7dbc2a test winograd is close to normal conv [pr] (#10557) 
Co-authored-by: chenyu <chenyu@fastmail.com>
 2025-06-06 19:11:49 -04:00
nimlgen
85cea23557 nv: original bw qmd (#10672) 
* nv: original bw qmd

* forgot
 2025-06-07 01:43:22 +03:00
George Hotz
5ef7c5923f docs: remove unused METAL_XCODE env var (#10421) 2025-06-06 18:39:54 -04:00
Sidharth N. Babu
ef14dfb277 compile fixes (#10442) 2025-06-06 18:38:37 -04:00
leopf
eb7305e6a4 Tensor.keccak("sha3_256") (#7186) 
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
Co-authored-by: George Hotz <geohot@gmail.com>
Co-authored-by: wozeparrot <wozeparrot@gmail.com>
 2025-06-06 15:24:05 -07:00
nimlgen
346b8542da nv: fix inval from gpu_get_id_info_v2 (#10670) 2025-06-07 00:54:32 +03:00
chenyu
bdede4924e fix odd number in get_test_global_size (#10671) 
factor might not be a integer if input global_size has an odd number in it
 2025-06-06 17:31:35 -04:00
George Hotz
bf4ffc054c mstack replaces scheduler complexity (#10654) 
* mstack replaces scheduler complexity

* leave that one

* contiguous

* work

* upd

* minimal failing test

* simpler

* attention is broken

* fix transformer

* failing tests

* real fix for llama

* kv cache test

* jit multi assign test

* better tests

* comment

* fix jit issue

* traverse after buf_uop
 2025-06-06 11:31:41 -07:00
George Hotz
7f0f97aa76 new test_multitensor tests (#10667) 
* new test_multitensor tests

* cleanup scheduler
 2025-06-06 10:26:28 -07:00
qazal
5170f387b3 remove UOp.metaop [pr] (#10664) 
* little simpler UOp.const_like [pr]

* remove UOp.metaop

* bind

* remove

* min diff

* that comment is fine
 2025-06-06 16:21:48 +03:00
chenyu
4a6d84c4c3 hotfix llama start_pos vmax is max_context-1 (#10659) 
* hotfix llama start_pos vmax is max_context-1

fixed `IGNORE_OOB=0 python3 examples/llama3.py --size 1B --benchmark --temperature 0`

* hotfix: multitensor transformer test tests kv cache

---------

Co-authored-by: George Hotz <geohot@gmail.com>
 2025-06-06 00:41:25 -04:00
George Hotz
5eb6e1e65a Revert "hotfix: multitensor transformer test tests kv cache" 
This reverts commit ad9f88419a.
 2025-06-05 21:15:34 -07:00
George Hotz
ad9f88419a hotfix: multitensor transformer test tests kv cache 2025-06-05 21:08:57 -07:00
George Hotz
8325c4f192 tests for multi assign (#10658) 
* tests for multi assign

* transformer tests

* add that assert
 2025-06-05 20:56:40 -07:00