geohot/tinygrad
1
0
Fork 0
forked from tinygrad/tinygrad
Code Pull Requests Actions Activity
13,406 Commits 4 Branches 15 Tags
942cb42b9703a48fdff19fa8051ec074ed5be006
Commit Graph
13406 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
geohot 942cb42b97 Revert "hotfix: bump Mac pytest timeout to 4 minutes" 
This reverts commit 695a0069ed.
 2026-05-25 17:25:11 -07:00
sirhcmandGitHub 8ddd1328df remove getenv(CI) (#16365) 
gone everywhere except test_interop, because torch MPS does not work in actions
 2026-05-25 20:23:33 -04:00
geohot 695a0069ed hotfix: bump Mac pytest timeout to 4 minutes 2026-05-25 17:20:19 -07:00
George HotzandGitHub 689ab6a49f move buffer view offset to src (#16364) 
* this work?

* failed
 2026-05-25 17:07:55 -07:00
sirhcmandGitHub d8f86be613 webgpu: shader-f16 support in arch (#16370) 2026-05-25 19:20:59 -04:00
qazalandGitHub 4bcc53eb26 viz: stable node position for +- toggle (#16367) 2026-05-26 06:30:47 +09:00
qazalandGitHub 3506eb08ec viz: sidebar toggles always recenter (#16366) 
* viz: sidebar toggles always recenters

* python brain
 2026-05-26 06:14:32 +09:00
chenyuandGitHub cdeb861828 invalids is empty [pr] (#16353) 2026-05-25 16:11:38 -04:00
qazalandGitHub b73d2d17b9 viz/cli: add --interval (#16363) 
* interval support

* add test_interval

* llama uses interval
 2026-05-26 03:35:06 +09:00
C TandGitHub 2ab90f31b1 use windows-specific alias nvcuda when loading cuda on windows (#16260) 
This also makes it possible to use cuda on windows by specifying 3 env
vars with direct dll paths: NVCUDA_PATH, NVRTC_PATH and NVJITLINK_PATH
without name collision with CUDA_PATH which is used for cuda headers
include path in NVRTCCompiler.
 2026-05-25 08:50:50 -07:00
wozeparrotandGitHub 68d2102fd2 llama: offload master weights (#16355) 2026-05-25 08:48:13 -07:00
qazalandGitHub eecd4706ff fix mailbox comment, add types (#16360) 2026-05-25 22:24:00 +09:00
nimlgenandGitHub 64095cf2e2 use get_buf in exec_kernel (#16356) 2026-05-25 15:13:40 +03:00
chenyuandGitHub 5d5e02871f remove Tensor.from_uop (#16344) 
and no device for const in Tensor init
 2026-05-24 18:53:09 -04:00
nimlgenandGitHub a891727c9f hcq2: multi (#16347) 
* hcq2: multi

* cleaner a bit
 2026-05-24 19:28:33 +03:00
chenyuandGitHub 926d125a63 update test_stack (#16345) 
also skip COMPILE_ONLY, it was comparing 0==0
 2026-05-23 10:42:35 -04:00
chenyuandGitHub 149a87dac2 deviceless const cleanups (#16341) 2026-05-22 20:11:01 -04:00
sirhcmandGitHub 35461d4d8f ci: cleanup some deps [pr] (#16340) 2026-05-22 19:16:08 -04:00
sirhcmandGitHub 451f38155c start cleanup of the slowest tests (#16339) 2026-05-22 18:39:36 -04:00
nimlgenandGitHub 26b3b3f6a2 hcq2: move submit lowering to schedule (#16330) 
* hcq: move submit lowering to schedule

* Dx
 2026-05-22 23:15:19 +03:00
wozeparrotandGitHub 2d48fe8b7b feat: bump version to 0.13.0 (#16337) v0.13.0 2026-05-22 13:12:45 -07:00
chenyuandGitHub acc519720b add missing init files, add chat.html to package-data (#16334) 2026-05-22 13:53:34 -04:00
googlefan256GitHubchenyu
 		eeadf26dad Fix no module named error (#16305) 
Co-authored-by: chenyu <[email protected]>
 2026-05-22 12:51:29 -04:00
nimlgenandGitHub 90dbb45563 nv: fix boot mem (#16332) 
* nv: fix boot mem

* linter
 2026-05-22 19:28:38 +03:00
nimlgenandGitHub 5d77a94923 am: mec_pipe0_reset on gfx12 only (#16331) 2026-05-22 19:02:18 +03:00
qazalandGitHub bbfe4f80ec quantize_fp8 kernels in uops (#16288) 
* add tests

* simple UOp kernel is n^2

* fast kernel matching c++, opts_to_apply=()

* remove cpp

* simple o(n) kernel, two passes

* fuse the loops

* works on DEV=CPU

* multi regression test

* fix multi, this can possibly be its own bugfix

* test cleanups

* minimal diff

* match C in UOps

* Revert "match C in UOps"

This reverts commit 0bef740c30.

* edit test

* match speed with C try 2

* needs_second_gpu

* cleanup
 2026-05-22 20:54:06 +09:00
chenyuandGitHub 3115952266 more unique const removal prerequisite (#16328) 2026-05-21 23:51:40 -04:00
sirhcmandGitHub c2d06570a5 remove getenv(CI) from core tinygrad (#16326) 2026-05-21 22:20:33 -04:00
chenyuandGitHub 9744d512d9 use more non-buffered const (#16327) 2026-05-21 21:37:52 -04:00
sirhcmandGitHub 150a82de1f start cleaning up dtype tests (#16324) 2026-05-21 21:11:49 -04:00
chenyuandGitHub 31424cda71 Tensor.requires_grad -> is_param (#16325) 
for optimizer
 2026-05-21 19:39:57 -04:00
sirhcmandGitHub 518e60534e only load tinymesa_cpu when LVP is explicitly requested (#16320) 2026-05-21 19:03:13 -04:00
chenyuandGitHub 720a27bed8 remove many requires_grad= args (#16321) 
* remove many requires_grad= args

* doc and example

* not cifar
 2026-05-21 18:37:11 -04:00
wozeparrotandGitHub 0c41317a59 llama: update 405b scripts (#16309) 2026-05-21 14:03:34 -07:00
wozeparrotandGitHub fb718a5e9d llama: realize amax (#16308) 2026-05-21 14:00:48 -07:00
chenyuandGitHub 73ea36f4ac full(buffer=True) (#16311) 
make full a buffer with flag to turn off
 2026-05-21 16:34:44 -04:00
George HotzandGitHub 6815f28849 dtype.vec shapes (#16287) 
* dtype.vec shapes

* something

* Closer

* more passes

* shape is in spec

* fix reduce

* image dtype shape correct

* lil

* use reshape on image

* need BUFFER there

* remove that test

* fix ptx + x86

* fix nir

* x86 fix maybe

* x86 fixups

* x86 fix

* don't check that for NOOP
 2026-05-21 11:56:49 -07:00
wozeparrotandGitHub afc5bfa183 llama: remove fused grad accum (#16301) 2026-05-21 09:38:40 -07:00
nimlgenandGitHub a321700baa hcq2: multi prereqs (#16304) 2026-05-21 17:00:52 +03:00
qazalandGitHub e33e058d34 set SPLIT_W13=0 for 8b DP by default (#16302) 2026-05-21 22:09:10 +09:00
sirhcmandGitHub dd279ee25e print dtype decomp warning in DEBUG=2 (#16300) 2026-05-20 22:08:48 -04:00
George HotzandGitHub ec547250ef don't use dtype vec for image idx (#16298) 
* don't use dtype vec for image idx

* double gate

* y/x confused

* upd

* fix nir

* simplify_valid_image_load
 2026-05-20 18:45:13 -07:00
sirhcmandGitHub 172f9493e1 move is_dtype_supported to renderer (#16226) 2026-05-20 21:19:37 -04:00
chenyuandGitHub d548f8d0f3 use clone instead of unique_const in allreduce [pr] (#16297) 2026-05-20 18:58:47 -04:00
qazalandGitHub 9e88b08f93 x86: don't use id (#16296) 
* x86: don't use id

* diff

* more minimal change

* unique
 2026-05-21 07:36:40 +09:00
sirhcmandGitHub da07b28998 am: override smu 13_0_7 to 13_0_0 (#16292) 2026-05-20 18:14:30 -04:00
chenyuandGitHub beea4633fc UOp.clone [pr] (#16295) 
generates the store after structure
 2026-05-20 17:47:49 -04:00
qazalandGitHub a19fa2908f fix x86 nondeterminism (#16293) 2026-05-21 05:48:05 +09:00
George HotzandGitHub 58d58c1659 remove DEVECTORIZE (#16290) 
* remove DEVECTORIZE

* fully remove DEVECTORIZE
 2026-05-20 13:25:49 -07:00
wozeparrotandGitHub 825f30bf18 llama: apply_grad saves memory (#16275) 2026-05-20 13:14:06 -07:00