github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-04-29 03:00:14 -04:00
Code Issues Packages Projects Releases Wiki Activity
12,959 Commits 248 Branches 14 Tags
7cbfa1896aebd6e1210d626db239fa54ba6d802c
Commit Graph

12959 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
chenyu
7cbfa1896a comment out unused arm, triton in toml (#15741) 
fixed `PYTHONPATH=. uv run tinygrad/apps/llm.py`
 2026-04-15 10:05:19 -04:00
Christopher Milan
1c36878008 DEV: suggest alternatives (#15732) 2026-04-14 23:42:32 -04:00
George Hotz
1ae6528bb6 move schedule into schedule (#15736) 
* move schedule into schedule

* callify to root

* sched docs
 2026-04-15 11:03:25 +08:00
wozeparrot
3721c60bef llama: bs 16 (#15737) 2026-04-14 19:52:03 -07:00
wozeparrot
480ad264a4 llama: per device amax (#15735) 2026-04-14 19:01:17 -07:00
Christopher Milan
adc96cd724 qcom: synchronize for copyin (#15731) 
fixes: #15698
 2026-04-14 18:31:15 -04:00
chenyu
3394d18066 size*itemsize -> nbytes (#15729) 
and some UOp.size removal to prep for size to mixin change
 2026-04-14 16:27:54 -04:00
nimlgen
e9ecc990ea amd: add r9700 devid (#15721) 2026-04-14 20:15:00 +03:00
George Hotz
2450c8cba8 rename to callify + fix mypy (#15727) 
* rename to callify + fix mypy

* update test
 2026-04-14 23:43:19 +08:00
chenyu
528faa18ec update env_vars.md (#15722) 
remove HCQ_VISIBLE_DEVICES, IMAGE=2 and old DEBUG=3 stuff
 2026-04-14 09:13:35 -04:00
George Hotz
359b1582d6 amd: EMU DPP support (#15719) 
* EMU DPP support from GPT 5.4

* cleanups

* simple

* nope

* fix
 2026-04-14 14:58:41 +08:00
wozeparrot
2b8d303f75 allreduce in precast dtype (#15689) 2026-04-13 20:24:12 -07:00
George Hotz
5683126844 llm: support for tekken tokenizer (#15720) 2026-04-14 10:52:07 +08:00
chenyu
70883a6950 cat the stack to mixin (#15715) 2026-04-13 18:44:39 -04:00
qazal
355e2729d3 viz: keep program UOp in data (#15714) 
* refactor program uop access

* c.name
 2026-04-14 07:04:16 +09:00
qazal
905b8adc97 viz: cli and server cleanups (#15713) 
* update get_profile arg[0]

* uop_to_json arg[0]

* data is standalone in cli
 2026-04-14 06:42:29 +09:00
Christopher Milan
d83707ec29 autogen: explicit types (#15679) 2026-04-13 16:54:39 -04:00
chenyu
ac41f15fc1 cumsum to mixin (#15712) 
built on top of getitem
 2026-04-13 15:06:08 -04:00
nimlgen
eac481b67f mlx: fix ctypes (#15711) 
* mlx: fix ctypes

* x
 2026-04-13 20:43:56 +03:00
nimlgen
b370f5c5ac hcq: call free for unmap (#15710) 2026-04-13 20:30:21 +03:00
chenyu
931d6cc62a basic getitem to mixin (#15697) 
* basic getitem to mixin

* cleanup

* fix

* cleanup
 2026-04-13 13:04:36 -04:00
George Hotz
7610bdc59e block multistore, it's not supported (#15708) 2026-04-13 20:57:59 +08:00
George Hotz
84d64b5835 hotfix: abstractions4 works in mock except asm 2026-04-13 20:57:00 +08:00
George Hotz
16f50a40a5 remove REMU from tree (#15706) 
* no more compare emulators

* remove remu from tree
 2026-04-13 20:43:08 +08:00
qazal
ac027055ef viz: no global state (#15705) 
* start viz data

* get_full_rewrites also moves

* update ref_map

* work

* update consumers

* cleaner cli

* linter

* cleanup tests

* back

* better

* sqtt tests
 2026-04-13 21:35:20 +09:00
George Hotz
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 Hotz
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 Hotz
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
b1tg
2b5ba0095d qwen3.5 (#15210) 
* qwen3.5

* faster

* or

* rm zero hack

* less float

* T=1

* clean

* clean

* 4b

* rope_dim

* Revert "jit: captures linears, not execitems (#15399)"

This reverts commit 9656d97d97.

* DeltaNetBlock

* pairwise_topk

* clean

* Reapply "jit: captures linears, not execitems (#15399)"

This reverts commit cf3deff53d.

* clean topk, _swiglu

* common

* FFNBlock

* clean

* half

* no mix

* qwen3.5 test

* fix ssm cache invalidation

* TransformerConfig

* SSMConfig

* clean

* reset_state

* llm: reuse server conversation tokens to avoid BPE roundtrip cache miss

* import error

* prefill

* none check

* put it back

* clean pairwise_topk

* symbolic: fold BIND(CONST, CONST) to CONST

* clean

* simpler pm

* _cached_msg_count

* stream decoder; ssm checkpoints

* rm checkpoint

* attn_output_gate

* conflict, attn_output_gate

* clean, less has_ssm, assert

* chunked prefill

* _reset_cache

* _reusable_prefix_len

* revert loop

---------

Co-authored-by: b1tg <b1tg@users.noreply.github.com>
Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
 2026-04-13 15:35:24 +08:00
qazal
2ada38f777 viz: execv after all producers complete (#15696) 2026-04-13 08:15:47 +09:00
chenyu
f7ff480fa6 start mixin getitem tests (#15695) 
goal is to make Tensor[idx].uop equal to Tensor.uop[idx]
 2026-04-12 18:54:33 -04:00
chenyu
77385ccb37 more trivial stuff to mixin (#15693) 2026-04-12 15:17:16 -04:00
chenyu
ff1de5ae13 normalize logsumexp contiguous_backward to mixin (#15692) 
* normalize logsumexp contiguous_backward to mixin

* more
 2026-04-12 13:13:00 -04:00
chenyu
0254cfe642 move usum and uprod to mixin (#15690) 
and used it to clean up ops and tensor
 2026-04-12 11:42:24 -04:00
nimlgen
e9b2e156b4 add jitbeam to tinygpu docs (#15691) 2026-04-12 18:20:26 +03:00
chenyu
e706f408cb suppress test warnings from numpy (#15688) 2026-04-11 22:33:20 -04:00
nimlgen
938cba4fdf amd: a bit faster usb, skip interrupts on sync (#15686) 2026-04-11 17:26:36 +03:00
qazal
054d78e6ff fix llama profile.sh NULL source (#15685) 2026-04-11 22:56:05 +09:00
Graham Robbins
4ca844e96b add Q1_0 gguf type (#15683) 
* add Q1_0

* better description

* fix trailing whitespace

---------

Co-authored-by: George Hotz <72895+geohot@users.noreply.github.com>
 2026-04-11 18:17:24 +08:00
George Hotz
5156a04cf5 add support for AM_POWER_LIMIT (#15684) 
* add support for AM_POWER_LIMIT

* level None
 2026-04-11 17:14:54 +08:00
wozeparrot
457508d5a0 llama: save more 2 (#15681) 2026-04-11 01:03:36 -07:00
George Hotz
29238b772f AMD USB: support for 0xF3 power toggle 2026-04-11 13:04:38 +08:00
George Hotz
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
wozeparrot
590464c8d8 llama: only support wqkv path + cleanups (#15680) 
* llama: only support wqkv path + cleanups

* llama: missing transpose
 2026-04-11 07:39:27 +08:00
nimlgen
aa012d6f08 usb: faster custom (#15678) 
* usb: _f0_out_buf for e4 cmd as well

* custom speed

* fast
 2026-04-10 23:00:31 +03:00
nimlgen
58646f9569 usb fast copyout (#15677) 
* usb

* fix usb
 2026-04-10 21:04:49 +03:00
qazal
0d5cdc9600 viz: split draw loop (#15676) 
* split draw loop

* one draw

* no functions

* inline all highlights

* cleanup
 2026-04-10 23:25:50 +09:00
chenyu
e1334d3852 move canonicalize_device to device.py (#15675) 2026-04-10 09:43:56 -04:00
chenyu
8e7fcc8ca3 remove _include_initial in _cumalu (#15674) 
handle negative pad in caller
 2026-04-10 08:33:30 -04:00
George Hotz
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