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

• cf2eea961c more beautiful_cartpole with exposed hparams geohot 2024-01-07 17:41:09 -08:00
• 44618427f1 add bf16 type_map for both cuda and hip (#3036) Yixiang GaoandGitHub 2024-01-07 14:26:55 -08:00
• ef5f545fd8 add more Tensor.clip test cases (#3034) chenyuandGitHub 2024-01-07 13:08:59 -05:00
• c9371f0d31 hotfix llama conversation mode (#3031) chenyuandGitHub 2024-01-06 16:57:07 -05:00
• fa707c81e5 move beautiful cartpole action sampling inside jit (#3028) chenyuandGitHub 2024-01-06 00:39:55 -05:00
• ebb81e8f11 hotfix: st.size() -> st.size in llama geohot 2024-01-05 20:18:52 -08:00
• a8ba1ac08f track size in shapetracker (#3026) George HotzandGitHub 2024-01-05 20:15:53 -08:00
• 138c17c094 enable argmax tests for METAL/WEBGPU in CI (#3027) chenyuandGitHub 2024-01-05 21:43:00 -05:00
• 2a2d3233d2 add test that the compiler isn't used (#3025) George HotzandGitHub 2024-01-05 17:24:01 -08:00
• 520406cf3a add Tensor.unflatten and Tensor.flatten(end_dim) (#3023) chenyuandGitHub 2024-01-05 17:55:29 -05:00
• f432ec9c33 Bitcast hip fix + fix mixtral (#3022) George HotzandGitHub 2024-01-05 14:51:25 -08:00
• eda43767de use Scalar = Union[float, int, bool] in tensor.py (#3021) chenyuandGitHub 2024-01-05 13:56:26 -05:00
• 60abc62a3f fast hip read (#3014) George HotzandGitHub 2024-01-05 10:33:13 -08:00
• 4465ef28c5 add test_softmax to test_ops (#3020) chenyuandGitHub 2024-01-05 11:19:49 -05:00
• 7c80b78be9 cleanup gpt2 build function (#3018) chenyuandGitHub 2024-01-04 23:14:53 -05:00
• 55e52abeba minor cleanup of matvec in hand_coded_optimizations (#3015) chenyuandGitHub 2024-01-04 19:43:49 -05:00
• f88506e630 move gpt2/llama sampling inside the model call (#3013) chenyuandGitHub 2024-01-04 17:01:50 -05:00
• c2a044ed83 disk_read_speed example geohot 2024-01-04 13:58:39 -08:00
• 8a63f26a0f make LR scheduler work with multigpu (#3011) Yixiang GaoandGitHub 2024-01-04 12:10:56 -08:00
• 8524493748 minor gpt2 cleanup (#3012) chenyuandGitHub 2024-01-04 13:53:18 -05:00
• 2b6670d2ea separate entry for HALF hlb_cifar10 in benchmark (#3010) chenyuandGitHub 2024-01-04 13:24:10 -05:00
• 5337211058 llvm CMPEQ chenyu 2024-01-04 10:01:12 -08:00
• b8c30eb358 no midcast MULACC for llvm chenyu 2024-01-04 09:53:12 -08:00
• 91665ef143 rewrite MUL CAST SUM to CAST MULACC chenyu 2024-01-04 09:42:22 -08:00
• ab7dfd637b use float for acc dtype for half tensor sum chenyu 2024-01-03 18:56:42 -08:00
• 6fa285b943 touchup onnx xor and not (#3008) chenyuandGitHub 2024-01-04 02:02:42 -05:00
• 57817028bb removed redundant dtype hacks in onnx_ops (#2939) 3 people 2024-01-04 14:45:24 +08:00
• 9f39165188 correct (dtype, device) in test_dtype.is_dtype_supported (#3007) chenyuandGitHub 2024-01-04 00:25:37 -05:00
• ae112c9dbe fix some long lines in tests (#3006) chenyuandGitHub 2024-01-03 23:53:33 -05:00
• 7e191fbb86 hotfix: don't jitcache with 1 kernel. improvements to hip sniffer geohot 2024-01-03 19:17:08 -08:00
• bcc1aa21ac make disk simpler (#3002) George HotzandGitHub 2024-01-03 17:46:21 -08:00
• 9699c8c90b don't alloc for InterpretedASTRunner (#2999) George HotzandGitHub 2024-01-03 17:05:53 -08:00
• bca0b95ee3 bump shapetracker simplify message to DEBUG >= 5 (#2998) chenyuandGitHub 2024-01-03 20:00:36 -05:00
• 74a30431b4 replace d[a] if a in d else b with d.get(a, b) (#2997) chenyuandGitHub 2024-01-03 18:10:25 -05:00
• 74cc6fd3c2 remove AndNode.__floordiv__ special case (#2996) chenyuandGitHub 2024-01-03 17:44:55 -05:00
• a0c7cb2564 hotfix: create weights dir in local tg checkout geohot 2024-01-03 14:14:33 -08:00
• fc36a7d669 tinygrad weights geohot 2024-01-03 14:09:28 -08:00
• 1ac4d27869 remove VariableOrNum from Node.substitute arg (#2995) chenyuandGitHub 2024-01-03 17:02:25 -05:00
• 65dc3700b7 hip device is default on supported platforms (#2993) George HotzandGitHub 2024-01-03 13:42:13 -08:00
• 77c98a1543 hotfix: remove weights directory geohot 2024-01-03 13:40:39 -08:00
• 0be0f2f745 remove stable diffusion test on tinymac geohot 2024-01-03 13:18:24 -08:00
• a354ec9dad Revert "hotfix: HIP is the default device on HIP platforms" geohot 2024-01-03 13:16:54 -08:00
• b748b569f5 hotfix: HIP is the default device on HIP platforms geohot 2024-01-03 13:13:52 -08:00
• 753a7ecc05 Hip driver (#2992) George HotzandGitHub 2024-01-03 12:53:47 -08:00
• f290ca3924 hotfix: save lines in graph geohot 2024-01-03 12:03:42 -08:00
• bc4b6e758b Merge pull request #2981 from g1y5x3/cifar_fp16 Yixiang GaoandGitHub 2024-01-03 11:15:42 -08:00
• d7d5a487ad hotfix: all device canonicalize should be done in Tensor geohot 2024-01-03 10:48:04 -08:00
• ea3bc2f509 remove wino benchmark for now Yixiang Gao 2024-01-03 10:46:43 -08:00
• 5663dd46b6 Merge branch 'master' of github.com:tinygrad/tinygrad into cifar_fp16 Yixiang Gao 2024-01-03 10:11:46 -08:00
• 81b97cd2c6 canonicalize device in LazyBuffer constructor (#2991) chenyuandGitHub 2024-01-03 12:55:25 -05:00
• db525cf8c2 multitensor failed test case with +1 then sum on DEVICE:0 (#2990) chenyuandGitHub 2024-01-03 12:17:11 -05:00
• 7f1802cd50 update benchmark Yixiang Gao 2024-01-03 09:09:34 -08:00
• 5dbaaa7061 hotfix: make multitensor shard contiguous geohot 2024-01-03 08:48:30 -08:00
• 590268fa03 out_tokens -> grouped in linearizer (#2989) chenyuandGitHub 2024-01-03 11:45:28 -05:00
• 8e1fd6ae9d test works Yixiang Gao 2024-01-03 07:22:01 -08:00
• 4f89f8b73a make sure the old hyp breaks the test Yixiang Gao 2024-01-03 07:13:54 -08:00
• 84eb6dd32a skip GPU cause opencl on intel can't compile half Yixiang Gao 2024-01-03 07:07:21 -08:00
• 73879b50ad only need to check the min_lr for the nan bug Yixiang Gao 2024-01-03 07:00:50 -08:00
• 99f8740c60 running half in CI CPU is slow Yixiang Gao 2024-01-02 18:44:35 -08:00
• 781690fd99 how long it takes on CI CPU without the lr scheduler Yixiang Gao 2024-01-02 18:33:48 -08:00
• dd00bcb9c0 fix whitespace Yixiang Gao 2024-01-02 18:16:33 -08:00
• 841487cad9 add half test with using hyp from benchmarks Yixiang Gao 2024-01-02 18:14:30 -08:00
• f494b9d463 simple multitensor API (#2903) George HotzandGitHub 2024-01-02 17:49:44 -08:00
• 5522ba234b simplify image functions (#2987) George HotzandGitHub 2024-01-02 17:35:08 -08:00
• 6e9406c986 one list comprehension in search action (#2988) chenyuandGitHub 2024-01-02 20:29:26 -05:00
• 08a34faea8 pass tuple for strs to startswith (#2986) chenyuandGitHub 2024-01-02 19:51:15 -05:00
• dbe4a1a914 switch CI to tiny8 (#2984) George HotzandGitHub 2024-01-02 16:40:25 -08:00
• b753d280f7 move hyp out of the train so it can be imported Yixiang Gao 2024-01-02 15:56:17 -08:00
• 0dd3ca59cd simpler ModNode.__mod__ and ModNode.__floordiv__ (#2983) chenyuandGitHub 2024-01-02 18:52:42 -05:00
• c07907e644 grad -> grad_output in mlops for consistency (#2982) chenyuandGitHub 2024-01-02 18:03:55 -05:00
• 54cdba57e7 mend Yixiang Gao 2024-01-02 14:21:06 -08:00
• 26303d181b re-enable half cifar benchmarks Yixiang Gao 2024-01-02 14:16:35 -08:00
• 2e4d9ad936 adjsut div factor to avoid underflow Yixiang Gao 2024-01-02 13:36:38 -08:00
• ad0d710ec4 merge apply_opt OptOps.LOCAL and OptOps.LASTLOCAL into one block (#2980) chenyuandGitHub 2024-01-02 16:40:10 -05:00
• 8de160d08e hotfix: remove dead code, save lines geohot 2024-01-02 12:52:20 -08:00
• 878e869663 simpler SumNode.__mod__ (#2979) chenyuandGitHub 2024-01-02 15:09:15 -05:00
• 91ddda244f minor cleanups in dtype.py (#2978) chenyuandGitHub 2024-01-02 13:42:37 -05:00
• ff5399f053 move one last dtype test from test_helpers to test_dtype (#2975) chenyuandGitHub 2024-01-02 12:37:56 -05:00
• deb3722aac refactor workitems (#2973) qazalandGitHub 2024-01-02 19:16:52 +02:00
• 01cdd6596f share hip and cuda (#2972) qazalandGitHub 2024-01-02 16:34:24 +02:00
• bd6a0c90a0 add Tensor.split (#2750) 3 people 2024-01-02 00:09:04 -06:00
• e7a432b479 search refactor (#2969) George HotzandGitHub 2024-01-01 17:39:26 -08:00
• b1d9e54ea3 regenerate kernel ast dataset (#2968) chenyuandGitHub 2024-01-01 20:26:17 -05:00
• cc2969f690 simpler cstyle (#2966) George HotzandGitHub 2024-01-01 16:20:10 -08:00
• 17f0c3006b hotfix: do stable diffusion first on mac geohot 2024-01-01 15:38:25 -08:00
• 58d3d5030b vars_from_ast -> LazyOp.vars (#2965) chenyuandGitHub 2024-01-01 18:12:38 -05:00
• 980f421442 hotfix: remove cast from beautiful_cartpole geohot 2024-01-01 15:02:03 -08:00
• a280cfe169 move dtypes to dtype.py (#2964) George HotzandGitHub 2024-01-01 14:58:48 -08:00
• fadaa2ec28 remove type check for LazyOp.src now it's always LazyOp (#2963) chenyuandGitHub 2024-01-01 17:27:29 -05:00
• c81ce9643d move globalcounters to ops (#2960) George HotzandGitHub 2024-01-01 14:21:02 -08:00
• 8291986959 Variable.sum -> Node.sum, Variable.ands -> Node.ands (#2961) chenyuandGitHub 2024-01-01 16:21:28 -05:00
• 3d720b5761 move expand_idx, iter_idxs and expand_node from symbolic to linearizer (#2959) chenyuandGitHub 2024-01-01 14:41:21 -05:00
• e0ecab3797 touchups from multibuffer branch (#2958) George HotzandGitHub 2024-01-01 11:33:41 -08:00
• 45247385eb hotfix: make the line counter correct geohot 2024-01-01 11:01:22 -08:00
• 56f44bd10e move the compiler cache to be global (#2957) George HotzandGitHub 2024-01-01 10:59:56 -08:00
• 063f465604 simpler webgpu (#2956) George HotzandGitHub 2024-01-01 10:28:59 -08:00
• fea20d71b3 add /opt/cuda/include directory (#2920) Shawn HaglerandGitHub 2023-12-30 10:16:42 -06:00
• 0d6e264c48 cleanup Tensor.triu and Tensor.tril (#2953) chenyuandGitHub 2023-12-29 22:27:18 -05:00
• e53b96fdbb fix TC=2 tensor core op test (#2951) chenyuandGitHub 2023-12-29 21:39:49 -05:00
• ad4472e6e8 cleanup llama apply_rotary_emb and other helpers (#2950) chenyuandGitHub 2023-12-29 11:39:15 -05:00