Commit Graph

Select branches

Hide Pull Requests

abstractions_4

add_local_buffer

after_qol

algebraic_upat

all_bv_no_idx

all_sink

amd_isel

amd_refac

amd_smu_reset_patch

amd_sqtt

amd_uop

amd_uop_cleanups

assign_copy

assign_try_3

before2

better_spec

better_syn

bitcast_spec

broadcast_shape_2

bt_ei

buf_heur

bump_nv_2404

cache_reshape

call_inline

call_is_kernels

callify

cast_vec_as_cast

cat_mop

changes_from_rangify

check_clsetarg

check_dims

cifar_30

cifar_30_clean

clean_load

clean_spec

clone_tg

codegen_try_2

contig_on_after

continue_mat

deepsek

default_late_allreduce

delete_kernel_py

delete_ones

delete_slow_rangeify

delta_attention

delta_attention_scan

demote_op

disable_rules

dsp_search

dsp_search_merged

dtype_shape

earlier_gater

emu_cdna

emulate_ctx

end_all_ranges

endrange

er_prioity

execution_unit

failing_assign

fancy_usb

fast_llm

fix_fa_fwd_clang_22

fix_fuse

fix_llvm_cast

fix_range_merging_sd

fix_rangeify_tests

fix_rc

fix_sched_copy

fix_slice_store

fix_src_spec

fixup_fix_kernel

full_spec

fun_w_egraphs

gate-on-load-store

gate_rewrite

gemma4_gpt

gen_pdf_fast

gh-pages

gl_dims

gldims

glm_flash

grad_outerworld

grok_x86_novec

group_div_rules

hcq2_vram_copies

heur_4096

heur_group

hk_gemm

http_threading

hwtest_fixes

id_store

if_range_cleanups

ignore_oob

image_idx_xy

image_in_mc

image_no_vec

image_refactors_stack

index_to_slice

invalid_try_3

jinja_template

kernel4

kernel_is_call

kernelless

kimi_slop

kitten_matmul

late_Locals

late_add_load

late_allreduce

late_codegen

late_decomp_is_late

late_lower_idx

late_view

late_weakint

lds_double_buffering

less_wmma_arg

lil_dg_change

lil_rangeify_td

llama_trainer

llm_262k_minimal

llm_flash_attn

llm_qwen

llm_serve_fixes

load_view_late

local_cache

locals_have_device

loop_end_op

loop_is_rng

lowerer_hacks

mac_pytest

manual_kernel_3

master

mega_lowerer

min_new_lin

min_outer_range

minor_views

mixins7

mlperf_training_v4.1

mlperf_training_v5.0

mlperf_training_v6.0

modern_scan

more_apps

more_early_comps

more_folding_rules

more_kernelcount

more_pcode

more_rangeify_pm_tests

more_ren_clean

more_uop_gc_test

more_uop_mm

more_vmap

move_devec

move_dtype_decomps

move_expander

move_gates_to_load_store

move_mop_cleanup

move_schedule

move_simp_views

moveleftright

mt_typing

multi_range

multioutput

new_ck

new_devec

new_invalid_and_image

new_linearizer

new_shape

new_test

no_assign_2

no_buf_0

no_decimals

no_late_simplify_marg

no_merge_ends

no_merge_views

no_mstack_in_allreduce

no_one_ranges

no_sip_usbgpu

no_tsink_base

no_tuplize

no_user_ck_contig

no_vec_x86_4

no_weakint_variable

noop_index

novecimg

num_batches_tracked_shape

olmo3_support

one_compiler

one_hcopt

only_assign_buffers

only_loop_backends

only_reg_emu

only_reg_emu2

outer_range

outerworld_work

pad_invalid_work

pad_mixin

param_call_is_function

parse_sqtt_more

pci_iommu

port_nir

postrange_hcopts

precompile_backward

prefill_qwen_fast

print_special_ops

ps_clean

pyremu_try

python_prg

python_speed

q4k_support

qcom-openpilot-13ms

qcom_830

qcom_mmapeak

qcom_speed_codex

qol_string

qwen36_27b_amd_900

qwen36_27b_amd_minimal

qwen36_amd_bugfixes

qwen_mergable

r_cleanups

r_new_opt

ranged_call

rangeify_2

rangeify_const_folding

rdna3_vibes

rdna4_asm

rdna4_gemm

real_pyrender

recurse_sub

reg_coalese

relu_pattern

remove_assign

remove_cast_hack

remove_ck_contig

remove_define_var

remove_forced_re

remove_on_stack

remove_programspec

remove_pylint

remove_slow_tests

remove_vec_2

remove_vec_3

renderers_use_shapes_2

replace_if_with_range

reprocess_node_experiments

reshape_trait

resnet-ci

revert-11547-fix_mismatch_red

revert-11715-revert-11711-amd_llvm_tc_cast

revert-11809-no_ki_in_gpudims

revert-11811-better_div_nesting

revert-11876-memplan_vbig

revert-12104-simplify_valid_in_symbolic

revert-12308-bufferize_early

revert-12731-jitlink_init_check

revert-12813-after_cleanups

revert-13046-better_cleanup_arange_buffers

revert-13311-prof_link_sqtt

revert-13583-rocr_desc_2

revert-14734-tload

revert-17141-full_rewrite_to_sink

revert-17408-_broadcasted-weak

revert-9438-simple_matmul_np_rand

revert_wrong_ck_fix

rewrite_rangeify2

rtoposort

scan_assign

scan_op_8

sched_cache

sched_lin

sd_load_simple

self_type

set_testpath

shrink_in_render

simp_spec

simple_gidxs

simple_pool

simple_priority

simple_rangeify

simpler-sd-compile

simpler_end

simpler_fusion

simpler_image

simpler_postrange

simpler_reduce

slice_like_shrink

spec_3

spec_cleanup

spec_ren

speed_one_linear

split_buf_idx

split_decomp_pass

split_sub

stack_0

store_is_void

string_ranges

support_opts_in_contig

swizzle_in_kernel

sym_llm

sym_work

symbolic-chunk-split

szdiff_bot_test

tc_is_heur

tensor_graph_single_sink

test_amd_crash

test_fa

test_from_postopt

test_hgemm

test_unit

tiny_scan

tmr

tool_call_support

topoprop

ttomsa-new_linearizer

tuplegtuple

two_stage_remove

typed_checks

typing_work

unused_permute_arg

uop_prg

upd_test4

update_benchmark

update_benchmark_search

update_mlperf

user_failure

vecless_x86_glm

verify_sink_dims_clean

vf_ish

view_after_const

view_in_codegen

viz_cbroadcast

warp_fun

wgpu-f16-packed

working_postopt

x86_moves

x86_numel

v0.10.0

v0.10.1

v0.10.2

v0.10.3

v0.11.0

v0.12.0

v0.13.0

v0.4.0

v0.5.0

v0.6.0

v0.7.0

v0.8.0

v0.9.0

v0.9.1

v0.9.2

Mono Color

• 4d8235d5f7 readme update geohot 2020-12-13 20:24:33 -08:00
• 1a1c63a08b Gan is real...Look what tiny just generated! (#192) NeuralLinkandGitHub 2020-12-14 09:53:12 +05:30
• 6785614239 tinygrad.utils to extra.utils fix in mnist_gan (#190) Marcel BischoffandGitHub 2020-12-12 23:52:36 -05:00
• d901ef6b23 🎉 effort to generate mnist data using GAN with tinygrad. [WIP] (#166) NeuralLinkandGitHub 2020-12-13 07:28:04 +05:30
• e6a5c6c93e Added indentation linter (#187) Mufeed VHandGitHub 2020-12-13 06:45:09 +05:30
• f95e79dab7 update readme geohot 2020-12-12 17:14:10 -08:00
• a5aced8d47 30 MEGAReLUs. we need to lose 12 lines geohot 2020-12-12 17:07:34 -08:00
• 49da969d25 Fixed a typo. (#189) WillemKaufandGitHub 2020-12-12 19:25:33 -05:00
• bc5df477de readme and .ane() geohot 2020-12-12 16:15:38 -08:00
• da873cd556 Single ReLU in ANE (#188) George HotzandGitHub 2020-12-12 16:11:34 -08:00
• 07ece2105e actually move it geohot 2020-12-12 15:26:58 -08:00
• 1d10559d1d tinygrad.utils -> extra.utils geohot 2020-12-12 15:26:07 -08:00
• 59358304a3 ane geohot 2020-12-12 15:23:21 -08:00
• 36d4eee323 fix compiler segfault geohot 2020-12-12 15:10:47 -08:00
• abb7b74208 relu in python geohot 2020-12-12 14:50:05 -08:00
• d3886035dd ane dylib geohot 2020-12-12 13:41:09 -08:00
• cf66d549c1 fix example ane geohot 2020-12-12 13:32:49 -08:00
• 566045cefc uint8 nope geohot 2020-12-12 13:13:29 -08:00
• 8c25431619 Faster but still general binop broadcasting (#159) 3 people 2020-12-12 12:19:46 -08:00
• bf9ba8718a Profile GPU and CPU copying. (#182) LiamandGitHub 2020-12-12 21:15:47 +01:00
• 8e8cbc74b3 Minor clean up (#184) James RobertsandGitHub 2020-12-12 00:25:29 +02:00
• f4faf401bc require_init_gpu() function selects GPU as device and falls back to CPU if none are available (#180) SkoshandGitHub 2020-12-11 19:21:59 +02:00
• c7e95ddb21 Add diamond model test (#181) DauletandGitHub 2020-12-11 17:21:36 +00:00
• 38b29f49dd abs (#172) Marcel BischoffandGitHub 2020-12-10 12:24:35 -05:00
• e79cda6dad Add pyopencl to dependency installs (#174) LiamandGitHub 2020-12-10 18:24:08 +01:00
• 8ab8a71d5d refactor (#178) NeuralLinkandGitHub 2020-12-10 22:53:36 +05:30
• d204f09316 some progress on batchnorms (draft) (#147) Marcel BischoffandGitHub 2020-12-10 01:14:27 -05:00
• 5d46df638a abs as non-first class operation using relu (#171) Marcel BischoffandGitHub 2020-12-09 15:20:34 -05:00
• 4c55c7208f no pow if mul will do geohot 2020-12-09 08:19:29 -08:00
• b85f17f247 more optim cleanup geohot 2020-12-09 08:18:10 -08:00
• 9a64d13b94 add conv biases and max pool geohot 2020-12-09 08:01:20 -08:00
• 99fa65f057 enable batchnorm in serious mnist geohot 2020-12-09 03:29:40 -08:00
• ffb96b2d0b batchnorm by marcelbischoff geohot 2020-12-09 03:23:04 -08:00
• 00e376f36c leaky relu as geohot suggested (#167) NeuralLinkandGitHub 2020-12-09 16:28:35 +05:30
• c225e62dd2 touchups geohot 2020-12-09 02:52:28 -08:00
• 89d0ff6989 Consistent testing (#137) LiamandGitHub 2020-12-09 11:25:27 +01:00
• 34b38dd4d0 Extra install requirements. (#164) LiamandGitHub 2020-12-09 11:22:47 +01:00
• 0e02f394ee serious_mnist geohot 2020-12-08 21:43:02 -08:00
• 24d688c184 win more lines for core library (#158) DauletandGitHub 2020-12-08 22:18:45 +00:00
• 9f77fd6135 🔨 refactor optim (#156) NeuralLinkandGitHub 2020-12-09 03:46:31 +05:30
• 4e1a0de392 fix rsub geohot 2020-12-08 10:05:21 -08:00
• c4540f1b8c Support scalars by kartik4949 geohot 2020-12-08 09:52:07 -08:00
• 97fd9c1237 zero_grad there to match readme geohot 2020-12-07 23:12:18 -08:00
• c63f950348 need zero grad now geohot 2020-12-07 23:10:43 -08:00
• b355cd2571 Mean axis (doesn't work) (#154) George HotzandGitHub 2020-12-07 22:58:34 -08:00
• 38f97c8c80 prepare for ops_ane geohot 2020-12-07 21:54:22 -08:00
• 7f249ec76d touch up geohot 2020-12-07 21:51:32 -08:00
• 58ccebd7cd Sum with axis (#153) Marcel BischoffandGitHub 2020-12-08 00:49:18 -05:00
• ac9fecb05d lots of notes geohot 2020-12-07 21:40:31 -08:00
• 8d1500f497 conv neuron geohot 2020-12-07 21:12:52 -08:00
• e4bb53b0e9 work out more geohot 2020-12-07 20:32:50 -08:00
• 4927ad1897 float16 weights in min.weights geohot 2020-12-07 20:15:15 -08:00
• 3aac9aefce fix GPU profiling geohot 2020-12-07 20:03:28 -08:00
• b2eca6d45f Format debug output (#152) James RobertsandGitHub 2020-12-08 00:07:14 +02:00
• c7973cb0a1 ugh buffer_np is bad geohot 2020-12-07 08:07:00 -08:00
• 088f280dc3 touchups geohot 2020-12-07 07:50:27 -08:00
• 0cf21881b7 hwx parse w/o macho mods geohot 2020-12-06 23:13:28 -08:00
• aa4161f63e use classmethods for Tensor helper funcs (#146) Josh SmithandGitHub 2020-12-07 01:35:43 -05:00
• e75a6d1b4b quadconv geohot 2020-12-06 20:39:50 -08:00
• 23664c99bd double conv geohot 2020-12-06 20:26:02 -08:00
• 1a0f826dc6 highlight the commands geohot 2020-12-06 20:03:21 -08:00
• 2f1f006003 we have docs geohot 2020-12-06 19:54:03 -08:00
• 3531e81f0e dumping ANE docs geohot 2020-12-06 18:58:36 -08:00
• 6e793e96c3 deeebug geohot 2020-12-06 17:49:17 -08:00
• dced0cb44b oops, path to weights geohot 2020-12-06 16:33:42 -08:00
• c57dc61ea7 simple op examples geohot 2020-12-06 16:32:26 -08:00
• 7babf38617 found concat geohot 2020-12-06 16:27:12 -08:00
• ddd6778423 add neuron geohot 2020-12-06 16:24:42 -08:00
• f2f2d6aea3 docs and noop geohot 2020-12-06 16:10:44 -08:00
• 6ba25834ee found some plists geohot 2020-12-06 15:57:28 -08:00
• e2184c20ad min weights, update golden geohot 2020-12-06 15:29:15 -08:00
• d4d8bd0337 make minimal plist for compare geohot 2020-12-06 15:10:15 -08:00
• 0845ec43c6 compile takes in plist geohot 2020-12-06 14:51:33 -08:00
• 00312b8ad1 batchnorm work geohot 2020-12-06 14:40:07 -08:00
• da514c2918 fix enet init geohot 2020-12-06 13:52:07 -08:00
• 3b982f2f7a get_parameters geohot 2020-12-06 13:47:28 -08:00
• 102e6356e9 replace layer_init_uniform with .uniform geohot 2020-12-06 13:44:31 -08:00
• c71a8ef222 remove unused Div op geohot 2020-12-06 13:02:12 -08:00
• 20f95de408 less lines geohot 2020-12-06 12:56:03 -08:00
• 629efb391f lose a few lines geohot 2020-12-06 12:42:41 -08:00
• 521098cc2f se optional, track time better geohot 2020-12-06 12:29:42 -08:00
• 609d11e699 trainer works with CIFAR geohot 2020-12-06 12:20:14 -08:00
• 80a9c777ba requires grad, optim in train enet geohot 2020-12-06 11:10:30 -08:00
• c66c27d22e get parameters geohot 2020-12-06 10:45:04 -08:00
• 51daaa43d4 fix memory leaks, add gc test geohot 2020-12-06 10:34:40 -08:00
• 1717daa859 reshape doesn't copy anymore geohot 2020-12-06 09:51:09 -08:00
• 62ee47fef8 add GPUBuffer class geohot 2020-12-06 09:45:13 -08:00
• 17659f7dd7 gpu speedup, tests work on M1 geohot 2020-12-06 09:05:49 -08:00
• b8deb36e56 train BS=16 for 32 steps geohot 2020-12-04 10:00:32 -08:00
• ad1b225722 oops, i broke fill geohot 2020-12-04 09:53:38 -08:00
• cb79c9838e make the GPU 25% faster by not recreating kernels geohot 2020-12-04 09:51:00 -08:00
• df81bf5985 it's the default, but make it explicit geohot 2020-12-04 09:43:41 -08:00
• 888689b57b proprotip geohot 2020-12-04 09:24:46 -08:00
• 2862b42bac install from github geohot 2020-12-04 09:06:25 -08:00
• 1290e01e2c all ops supported on GPU now George Hotz 2020-12-03 10:43:11 -08:00
• 621a93b777 ane in readme George Hotz 2020-12-03 10:40:31 -08:00
• 1dcaecacc4 Support for Apple Neural Engine (#130) George HotzandGitHub 2020-12-03 10:32:26 -08:00
• c83cebccda Made the readme more consistent (#136) baplouandGitHub 2020-11-28 09:20:02 -05:00
• 541330c42a Update README.md (#133) Marcel BischoffandGitHub 2020-11-25 10:53:54 -05:00
• 0bbf66627c Define ProfileOp class once (#131) Mufeed VHandGitHub 2020-11-25 09:09:13 +05:30