Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
423 commits
Select commit Hold shift + click to select a range
68df797
Import Fix
lucylq Aug 12, 2025
f478800
Fix pyre errors
izaitsevfb Aug 12, 2025
5ab8f52
[Backend Tester] Add LSTM tests (#13238)
GregoryComer Aug 12, 2025
6157370
[Backend Tester] Add avgpool tests (#13239)
GregoryComer Aug 12, 2025
fa059f2
[Backend Tester] Add maxpool tests (#13240)
GregoryComer Aug 12, 2025
eeaaad6
[Backend Tester] Add adaptive avgpool tests (#13241)
GregoryComer Aug 12, 2025
04689e3
[Backend Tester] Add adaptive maxpool tests (#13242)
GregoryComer Aug 12, 2025
732d9de
[Backend Tester] Add Qualcomm tester and register flow (#12739)
GregoryComer Aug 12, 2025
5e05ca8
[Backend Tester] Add CSV report generation (#12741)
GregoryComer Aug 12, 2025
252f6d3
Update build from source docs (#13210)
GregoryComer Aug 12, 2025
da3a5e5
Move to Span<EValue*> instead of EValue** in delegate interface
JacobSzwejbka Aug 12, 2025
f86285c
Update XNNPACK to 3131afe (#13234)
mcr229 Aug 12, 2025
3c9e77b
[Backend Tester] Add tensor error statistic reporting (#12809)
GregoryComer Aug 12, 2025
6bd15dd
[Backend Tester] Report quantization and lowering times (#12838)
GregoryComer Aug 12, 2025
fbba9c2
Added ETDump to Wasm (#13304)
Conarnar Aug 12, 2025
55b6e74
[Backend Tester] Report delegation statistics (#12846)
GregoryComer Aug 12, 2025
172539b
[Backend Tester] Report PTE size (#13249)
GregoryComer Aug 12, 2025
9b3b270
[ET-VK] Add optional blocklist and allowlist to vulkan partitioner to…
SS-JIA Aug 12, 2025
5692111
[ET-VK][ez] Fix `vulkan_schema` interface and link Vulkan to pybindin…
SS-JIA Aug 12, 2025
acdc2e4
[ET-VK][examples] Create export script for Vulkan examples (#13294)
SS-JIA Aug 12, 2025
bb444f8
[ET-VK][CI] Add vulkan CI to test exporting and running models (#13295)
SS-JIA Aug 12, 2025
ab9eea8
Add csv and etdump files to gitignore (#13301)
mcr229 Aug 12, 2025
da39ab7
[Backend Tester] Add portable test flow (#13250)
GregoryComer Aug 12, 2025
7989804
Static attention IO manager: fix causal mask bug for last input position
sxu Aug 12, 2025
8cfb917
Update buck srcs for tester harness
GregoryComer Aug 12, 2025
78ceda2
Check the number of inputs in Method::set_inputs (#13341)
shoumikhin Aug 13, 2025
1a67f6c
Add TorchAO wrapper config to allow filter_fn for quantize_ (#13264)
abhinaykukkadapu Aug 13, 2025
c275f40
Add coreml quant recipes (#13265)
abhinaykukkadapu Aug 13, 2025
400f31a
Fix 8w8a qat qconfig setting activations
navsud Aug 13, 2025
07f8a4f
NXP backend: Return number of Neutron nodes in exception (#13126)
StrycekSimon Aug 13, 2025
4f3b0e7
Enable strongly typed ops for deployment
mcremon-meta Aug 13, 2025
353aff0
Arm backend: Run pytest model tests in parallell (#13322)
Erik-Lundell Aug 13, 2025
012af23
Arm backend: Update arm_tester.py with generate_etrecord (#13370)
perheld Aug 13, 2025
d0cf666
Arm backend: Add logit decomposition pass and test (#13366)
tbergkvist Aug 13, 2025
51efa78
Arm backend: Initial int16 extension (#13318)
per Aug 13, 2025
587aee8
Arm backend: Remove submodule serialization_lib (#13182)
zingo Aug 13, 2025
fcd9538
Arm backend: Enable semihosting in build_executor_runner.sh (#13325)
Erik-Lundell Aug 13, 2025
13eb90d
Fix torchao deps (#13107)
metascroy Aug 13, 2025
9675ed7
Revert "Add TorchAO wrapper config to allow filter_fn for quantize_ (…
swolchok Aug 13, 2025
ea8331b
Introduce set input/output API to Module. (#13363)
shoumikhin Aug 13, 2025
9493377
Update Llama Example Readme (#12956)
anzr299 Aug 13, 2025
c69af4c
Support for Running Arm Zephyr Toolchain on aarch64 Host Machines (#1…
BujSet Aug 13, 2025
da601a8
[ET-VK] Better work group sizes for matmul (#13378)
pytorchbot Aug 13, 2025
33176f9
[ET-VK] Add mechanism to trigger command buffer re-encode only when n…
pytorchbot Aug 13, 2025
caf61ed
Introduce unload method API to Module. (#13364)
shoumikhin Aug 13, 2025
4b6dfa0
[Backend Tester] Clean up a few test issues (#13258)
GregoryComer Aug 13, 2025
2404756
[Backend Tester] Clean up report output (#13306)
GregoryComer Aug 13, 2025
db3fd27
[Backend Tester] Write report progressively (#13308)
GregoryComer Aug 13, 2025
d14d4c5
create and validate build_variables.bzl (#8326)
swolchok Aug 13, 2025
bc8a57f
[Backend Tester] Add subtest index field (#13311)
GregoryComer Aug 13, 2025
4800771
[Backend Tester] Reduce log verbosity / spam (#13312)
GregoryComer Aug 13, 2025
5983be9
NXP Backend: Add infrastructure for pre processing passes in edge dia…
MartinPavella Aug 13, 2025
b36b71a
Android preset (#11119)
kirklandsign Aug 13, 2025
dcdf9db
Qualcomm AI Engine Direct - Static Decoder Runner Support 16bit KV IO…
winskuo-quic Aug 13, 2025
5a88920
Qualcomm AI Engine Direct - Phase out QCIR flow since it is no longer…
DannyYuyang-quic Aug 13, 2025
ee111e0
Bump Zephyr SDK Version in CI Image from v0.16.0 -> v0.17.2 (#13380)
BujSet Aug 13, 2025
69ba8e9
[Backend Tester] Seed based on test name (#13313)
GregoryComer Aug 13, 2025
58cfa13
Add support for strongly typed op_quantized_linear_out
ethansfng Aug 13, 2025
1f1cc08
removed cron schedule runs to workflow until the GitHub token issue i…
nil-is-all Aug 13, 2025
15b51ce
Set a doc build variable for executorch version (#13351)
kirklandsign Aug 13, 2025
b9ebbe7
lintrunner -a backends/qualcomm/CMakeLists.txt (#13396)
swolchok Aug 13, 2025
39fd4b7
Update the comments of calculate_numeric_gap
Gasoonjia Aug 14, 2025
a64aa44
Use dtype agnostic op_cat implementation, add op_cat testcases
ethansfng Aug 14, 2025
dd69066
Arm backend: Allocate buffers with alignment (#13412)
Erik-Lundell Aug 14, 2025
4c0f087
NXP backend: Improve cifarnet speed by removing the initial pading. (…
jirioc Aug 14, 2025
4565979
Arm backend: Generate ETRecord from arm_aot_compiler (#13273)
zingo Aug 14, 2025
cc10e16
Use unlifted export pass to tag delegated constants (#13407)
pytorchbot Aug 14, 2025
9bb3bbc
Add support for strongly typed op_quantized_relu (#13345)
ethansfng Aug 14, 2025
bd8f812
[Backend Tester] Add test flow CLI arg (#13360)
GregoryComer Aug 14, 2025
1fec15c
Ensure the correct output data type for the full op.
eigen-k Aug 14, 2025
a61bb5a
Arm backend: Add decomposition and test for acos (#13414)
emmakujala Aug 14, 2025
347afd1
Arm backend: Introduce documentation for VGF (#13369)
robell Aug 14, 2025
9729c5d
Update ExecuTorchValue.mm (#13425)
shoumikhin Aug 14, 2025
dcecaab
Remove outdated comments and names in OSS
mcremon-meta Aug 14, 2025
d286910
Fix typo in op_quantized_relu_asym8u_asym8u
ethansfng Aug 14, 2025
6627cbc
Fix unused-local-typedef issue
r-barnes Aug 14, 2025
a48dbfc
Refactoring Portable Operators to Standardize op_name Format (#12941)
BujSet Aug 14, 2025
180baba
Android scheduled build add vulkan (#13428)
kirklandsign Aug 14, 2025
1fdbafe
NXP Backend: Buck fixes for the PassManager
digantdesai Aug 14, 2025
5f1dd11
Add support for strongly typed op_quantized_matmul, generalize dispat…
ethansfng Aug 14, 2025
0d1c3dc
Arm backend: use tosa_ref_model only if installed
digantdesai Aug 14, 2025
6421fd3
[executorch] Add TorchAO wrapper config to allow filter_fn for quanti…
pytorchbot Aug 14, 2025
bf7de85
[executorch] Add coreml quant recipes (#13441)
pytorchbot Aug 14, 2025
0e90935
Static attention: do not specialize on input sequence length
sxu Aug 14, 2025
b4b1ac5
Fix typo in target file (#13443)
abhinaykukkadapu Aug 15, 2025
de54277
forward fix
cccclai Aug 15, 2025
85b9577
Switch to conda-forge on MacOS (#13442)
huydhn Aug 15, 2025
6be925a
use dtype agnostic implementation for non optimized op_permute_copy
ethansfng Aug 15, 2025
881bd12
Extend `PyBundledModule` with `extension.BundledModule`
Gasoonjia Aug 15, 2025
06366c5
Improve optimum coverage in ET (more models, xnnpack on mac) (#13400)
jackzhxng Aug 15, 2025
e489db3
build_variables.bzl: split PROGRAM_NO_PRIM_OPS_SRCS from EXECUTORCH_C…
swolchok Aug 15, 2025
1221ace
Split quantized convolutions into NCHW and NHWC variants
mcremon-meta Aug 15, 2025
9439d8a
[ez] Fix idx in duplicate_constant_node_pass (#13461)
SS-JIA Aug 15, 2025
6b7001d
build_variables.bzl: split PLATFORM_SRCS from EXECUTORCH_CORE_SRCS (#…
swolchok Aug 15, 2025
d93e407
Arm backend: Add missing using-declaration in VGFBackend.cpp (#13460)
YufengShi-dudu Aug 15, 2025
605b10c
[EZ] Replace `pytorch-labs` with `meta-pytorch`
ZainRizvi Aug 15, 2025
0b6e14a
Allow for HOP to be in the etreord graph
kimishpatel Aug 15, 2025
e8b1082
build_variables.bzl: split PATTERN_SRCS from PORTABLE_KERNELS_SRCS (#…
swolchok Aug 15, 2025
2c9436d
build_variables.bzl: make THREADPOOL_SRCS usable in BUCK (#13354)
swolchok Aug 15, 2025
7dae813
build_variables.bzl: make MPS_BACKEND_SRCS usable in BUCK (#13355)
swolchok Aug 15, 2025
ccb00ee
build_variables.bzl: make XNNPACK_BACKEND_SRCS usable in BUCK (#13356)
swolchok Aug 15, 2025
b5c558b
build_variables.bzl: make CUSTOM_OPS_SRCS usable in BUCK (#13357)
swolchok Aug 15, 2025
4f6b029
Expose portable ops as utils (add/stack)
manuelcandales Aug 15, 2025
d5d91bc
Introduce out shape utils (add/stack) (#13199)
manuelcandales Aug 15, 2025
6d00d37
switch top-level ExecuTorch build from executorch_srcs.cmake to build…
swolchok Aug 15, 2025
724dcb1
Alternative format specifier for %zd
cmt0 Aug 15, 2025
b8ab343
[ET-VK] Move rotary embedding custom op to be handled via graph pass …
pytorchbot Aug 16, 2025
dc7f9de
Qualcomm AI Engine Direct - Static LLM Decoder Refactor (#13314)
winskuo-quic Aug 16, 2025
9a5af57
[ET-VK] Enable IntxWeightOnlyConfig (#13466)
pytorchbot Aug 16, 2025
ed4e59f
Add aten::_upsample_bilinear2d_aa.out (#13458)
mergennachin Aug 16, 2025
29256b0
Add support for strongly typed conv_nchw and conv_nhwc
ethansfng Aug 16, 2025
a06b3da
Arm backend: Add dim_order_ops:: to the auto gen_oplist generations (…
zingo Aug 18, 2025
3d30f7f
Arm backend: Replace .export_for_training with .export (#13280)
AdrianLundell Aug 18, 2025
fce39c0
Arm backend: Move TOSA operators to dialect (#13408)
per Aug 18, 2025
6c506ae
Arm backend: Add example linkerscripts for U55/U85 (#13404)
perheld Aug 18, 2025
d99c9d2
Arm backend: Add cumsum support (#13457)
AdrianLundell Aug 18, 2025
892db7a
Arm backend: Add partial vulkan runtime support for VgfPipeline (#13471)
YufengShi-dudu Aug 18, 2025
fc25fd8
NXP backend: Remove optimization in fuse_quanitze_into_preceding_ops.…
roman-janik-nxp Aug 18, 2025
543a3c5
NXP backend: Remove optimization in prune_cast_operators.py (#13377)
roman-janik-nxp Aug 18, 2025
af656dc
NXP backend: Remove optimization in prune_reshape_operators.py (#13413)
roman-janik-nxp Aug 18, 2025
4f4c34b
Remove outdated NCHW to NHWC pass and rename the current one to Repla…
mcremon-meta Aug 18, 2025
d210198
NXP backend: Improve target support checks. (#13367)
MartinPavella Aug 18, 2025
f287e0a
Add a default image prefiller implementation
larryliu0820 Aug 18, 2025
1680283
Switch non-top-level ExecuTorch builds (size test, examples, etc.) fr…
swolchok Aug 18, 2025
9e38ee1
Enable BNNS copy for FP16 to FP32
metascroy Aug 18, 2025
e1e3933
Fix test-binary-size-linux -Wsign-compare failure with c10::irange (#…
swolchok Aug 18, 2025
455071c
Arm backend: Update tosa dialect buck file
digantdesai Aug 18, 2025
0d039c9
[ET-VK][ez] Move execute node threshold calculation from `prepare_pip…
pytorchbot Aug 19, 2025
259aa8b
[ET-VK] Runtime support for NamedDataMap (#13498)
pytorchbot Aug 19, 2025
bc5d91c
[ET-VK][AOT] Serialize constant tensors via NamedDataMap (#13499)
pytorchbot Aug 19, 2025
2b7e058
[ET-VK] Allocate memory for weight and activation tensors lazily (#13…
pytorchbot Aug 19, 2025
6b5f73b
[ET-VK][ez] Fix erroneous cherry-pick bot merge (#13512)
SS-JIA Aug 19, 2025
0c86282
Arm backend: Add support for QAT+per-channel combo (#13511)
oscarandersson8218 Aug 19, 2025
93eb208
Reset Temp Allocator after each use
cmt0 Aug 19, 2025
be3b509
Stop validating that build_variables.bzl matches buck-generated execu…
swolchok Aug 19, 2025
b0d5391
Only support int8 and quant dtypes for quant operators (#11685)
aaron-ang Aug 19, 2025
fc00827
Stop looking for buck2 in the top-level ExecuTorch build (#13393)
swolchok Aug 19, 2025
9d64ccf
Summary: Add MCU model script to validate and run the models (#13439)
psiddh Aug 19, 2025
293072c
Add a generic multimodal runner (#13166)
larryliu0820 Aug 19, 2025
d85205d
Force -O3 for executorch op_div.cpp in clang 19
akrieger Aug 19, 2025
55dfc90
Remove NTSTATUS cast
SamGondelman Aug 19, 2025
ae6d536
Clean up apparently-unnecessary mentions of BUCK2 in scripts (#13394)
swolchok Aug 19, 2025
38b86a3
Qwen and Phi-4-Mini targets (#13449)
rohansjoshi Aug 19, 2025
8f286f3
Wrap LLM runner tests in anonymous namespace (unbreak unittest-releas…
swolchok Aug 19, 2025
2bc6b0d
Delete extract_sources.py and cmake_deps.toml (#13395)
swolchok Aug 19, 2025
171451b
[Backend Tester] Add nightly CI job for XNNPACK (#13390)
GregoryComer Aug 19, 2025
09a4511
Add Model Profiling Automation Script (#13493)
leafs1 Aug 19, 2025
ee9f94c
Remove unused PROGRAM_SCHEMA_SRCS from build_variables.bzl (#13432)
swolchok Aug 19, 2025
f621114
removed lines of cron schedule runs until the github token issue is f…
nil-is-all Aug 19, 2025
4313e48
[Backend Tester] Mark adaptive avgpool2d as an unsupported portable o…
GregoryComer Aug 19, 2025
02a4657
[Backend Tester] Run Vulkan tests in nightly CI (#13445)
GregoryComer Aug 19, 2025
86c9ee1
[Backend Tester] Run Core ML tests in nightly CI (#13446)
GregoryComer Aug 19, 2025
4174b03
Improve softmax perf when transpose is not needed
mcremon-meta Aug 19, 2025
a08eb08
Remove other extensions' source files from EXTENSION_TRAINING_SRCS an…
swolchok Aug 19, 2025
f013ba4
Fix `buck query //extension/flat_tensor:` in OSS (#13484)
swolchok Aug 19, 2025
38ba8cf
Qualcomm AI Engine Direct - Remove input_list dependencies (#13411)
chenweng-quic Aug 19, 2025
faff634
Qualcomm AI Engine Direct - GA Static Smollm2 (#13406)
chenweng-quic Aug 19, 2025
c06b947
Remove pinning to ET commit for Zephyr CI job (#13388)
BujSet Aug 20, 2025
58ddf4f
Refactor pybinding unit test
cccclai Aug 20, 2025
fe255a2
Update project.pbxproj (#13537)
shoumikhin Aug 20, 2025
4797f2e
Fix Olmo trunk test and skip XNNPack trunk tests on mac (#13528)
jackzhxng Aug 20, 2025
ab4fd57
Re-enable model tests with recipes for xnnpack backend (#13519)
abhinaykukkadapu Aug 20, 2025
4c084e8
QNN Llama Runner implement IRunner (#13171)
rohansjoshi Aug 20, 2025
7d13b2e
Arm backend: Remove get_output_nodes from runner_utils. (#13417)
AdrianLundell Aug 20, 2025
88642c0
Arm backend: Fix for combo neg(x)+1 + tests (#13517)
wwwind Aug 20, 2025
2eb7d7e
Adding smollm2 to examples/models/__init__.py (#13514)
SaoirseARM Aug 20, 2025
ba6a40c
Arm backend: Add limited support for fish shell in setup_path.sh (#13…
zingo Aug 20, 2025
1f63c65
Arm backend: Update examples/arm/README.md (#13546)
zingo Aug 20, 2025
58b3199
Arm backend: Added VGF minimal example (#13545)
SaoirseARM Aug 20, 2025
c5a8e8f
Temporarily disable test-models-arm-zephyr (#13548)
digantdesai Aug 20, 2025
8b3261f
Fix buck cquery //extension/llm/runner: in OSS (#13527)
swolchok Aug 20, 2025
075988d
Remove unused sleef.h which breaks cross-compilation on windows (#13079)
ykhrustalev Aug 20, 2025
e9754ab
Bump the PyTorch pin to 20250811 (#13334)
swolchok Aug 20, 2025
af9f8a5
Link QNN backend to pybinding lib when built (#13467)
GregoryComer Aug 20, 2025
c9f0159
Split on dilation for strongly typed convs
ethansfng Aug 20, 2025
5da6516
Fix sigmoid operator to support boolean tensor inputs (#13515)
notkisk Aug 20, 2025
19ea8a6
Fix method meta error in wasm build (#13496)
Conarnar Aug 20, 2025
9ed517c
[Backend Tester] Add test flows for QNN (#13469)
GregoryComer Aug 20, 2025
cb4eeb4
[Backend Tester] Add additional quantized test flows for XNNPACK and …
GregoryComer Aug 20, 2025
5c44446
[Backend Tester] Add markdown summary in CI (#13535)
GregoryComer Aug 20, 2025
d638703
Add support for strongly typed quantized_op_add
ethansfng Aug 21, 2025
4019da4
Whisper audio processor
rohansjoshi Aug 21, 2025
454b8a1
Non-fatal error when ET_SWITCH encounters unsupported dtype
manuelcandales Aug 21, 2025
6fc8ede
Update test_remove_unused_parameters_pass.py (#13563)
digantdesai Aug 21, 2025
cc0609b
Fix memory planning for greed with heuristic algo.
hsharma35 Aug 21, 2025
45765ae
Input position accessor for static attention IO manager
sxu Aug 21, 2025
f73d44d
Unbreak build-benchmark-app (apple) after pin bump in #13334 (#13582)
swolchok Aug 21, 2025
65dc152
Summary: Add ExecutorchRuntimeException: Throw relevant exceptions f…
psiddh Aug 22, 2025
7616da9
Update tokenizer to include tekken implementation (#13601)
mergennachin Aug 22, 2025
39afccc
Add missing backslashes in example run section (#13603)
Conarnar Aug 22, 2025
e610f23
Make IOManager use Module instead of Method. (#13542)
shoumikhin Aug 22, 2025
391ae3c
Access Method directly from TrainingModule. (#13602)
shoumikhin Aug 22, 2025
99b4216
Make TensorPtr constructor check the data dize matches the shape. (#1…
shoumikhin Aug 22, 2025
5997ee3
Add set_outputs() API. (#13609)
shoumikhin Aug 22, 2025
80adad5
Create stale.yml workflow to label stale PRs (#13565)
nil-is-all Aug 22, 2025
619bc30
Add get_output API. (#13610)
shoumikhin Aug 22, 2025
20e60bf
Set an empty EValue input for models that expect None arg. (#13621)
shoumikhin Aug 23, 2025
bdea7d0
Split on depthwise for strongly typed convs
ethansfng Aug 23, 2025
88588bf
migrate all test_aten_ops to facto
zonglinpeng Aug 23, 2025
9c0280c
fix MM nullptr from zero bias
zonglinpeng Aug 23, 2025
3bb031b
Call .detach() in static attention cache update helper
sxu Aug 23, 2025
64d88aa
Event Tracer Constraint
cmt0 Aug 23, 2025
0d4fd84
[ET-VK][ez] Fix partitioner logic of finding keepdim arg of reduce op…
SS-JIA Aug 23, 2025
bb0ec6e
[ET-VK][ez] Support grouped convolutions (#13599)
SS-JIA Aug 23, 2025
e98da6d
[ET-VK][ez] Use XNNPACK's FuseBatchNorm pass (#13600)
SS-JIA Aug 23, 2025
67b48c3
[ET-VK][testing] Add scripts to facilitate operator testiing (#13593)
SS-JIA Aug 23, 2025
fb99e23
[ET-VK][ez] Consolidate tensor metadata calculation + buffer binding …
SS-JIA Aug 23, 2025
3a8edfd
[ET-VK] Introduce `BufferMetadata` GLSL struct to abstract tensor lay…
SS-JIA Aug 23, 2025
1653dbf
[ET-VK][ez] Allow high dimensional tensors (for buffer storage) (#13596)
SS-JIA Aug 23, 2025
4f7871a
[ET-VK] High dim tensor support for view, unsqueeze, squeeze, clone (…
SS-JIA Aug 23, 2025
c91401e
[Core ML] Improve asset management (#13560)
cymbalrush Aug 23, 2025
6142858
Rename stale to stale.yml (#13619)
nil-is-all Aug 23, 2025
4298ff1
fix mismatch sub dtype (#13447)
cccclai Aug 23, 2025
50b7913
NXP backend: Add support for the `aten.cat` operator. (#13505)
MartinPavella Aug 24, 2025
044bdcd
NXP backend: Add implementation of Tanh operator converter (#13510)
MartinPavella Aug 24, 2025
6f05c35
Arm backend: Dont try to fuse const for TOSA ops (#13575)
per Aug 25, 2025
9b7c80a
NXP backend: Fix `tanh` merge conflict. (#13626)
MartinPavella Aug 25, 2025
1dba47f
Cortex_m backend: Loosen edge op check. (#13550)
Erik-Lundell Aug 25, 2025
bf2abab
NXP backend: Use zero point for quantized padding. (#13576)
MartinPavella Aug 25, 2025
4c510f1
Fix aten.amax lowering issue (#13381)
cccclai Aug 25, 2025
f55769d
Update coremltools to 9b1 (#13614)
metascroy Aug 25, 2025
7854fe7
Add check_for_installed_private_headers_in_cmake_out (#13485)
swolchok Aug 25, 2025
4e316d4
NXP backend: Add support for conversion of Conv1D operator (#13549)
roman-janik-nxp Aug 25, 2025
f9593d2
Update cpuinfo pin to latest (#13624)
pytorchbot Aug 25, 2025
290a8f5
Added JS bindings for tokenizers library (#13566)
Conarnar Aug 25, 2025
ed11370
Run all periodic models when ciflow/periodic label is present (#13634)
shoumikhin Aug 25, 2025
3c84d53
Add support for data path in iOS (#13620)
lucylq Aug 25, 2025
fbff62e
Summary: Follow up fix to pr#13526 (#13640)
psiddh Aug 25, 2025
fd921c4
Allow none and string input types for Method (#13645)
shoumikhin Aug 25, 2025
ecf5be2
Resurface low level runtime API page (#13651)
mergennachin Aug 25, 2025
315c837
Fully enable the stale PR workflow (#13656)
nil-is-all Aug 25, 2025
be22ad5
Qualcomm AI Engine Direct - Scripts and accuracy improvement for Qwen…
winskuo-quic Aug 25, 2025
dc4ff25
Qualcomm AI Engine Direct - Improve GA Static Phi-4-mini accuracy (#1…
shewu-quic Aug 25, 2025
c72accb
Qualcomm AI Engine Direct - Fix broken unpacking in T5 dataset loadin…
DannyYuyang-quic Aug 25, 2025
1797ba1
Fix error reporting in Windows preset build job (#13247)
GregoryComer Aug 25, 2025
1feb7c7
Fix devtools CMake build failure on Windows (#13251)
GregoryComer Aug 25, 2025
ceb6f32
Create Windows CMake preset (#13257)
GregoryComer Aug 25, 2025
b78e768
Temporarily disable windows preset build in CI (#13669)
GregoryComer Aug 25, 2025
68a9a42
Increase binary size limit by 8 bytes (#13671)
shoumikhin Aug 26, 2025
c92def6
NXP backend: Remove IR optimization to remove dead branches. (#13574)
MartinPavella Aug 26, 2025
80d1407
Inline requantize kernels
mcremon-meta Aug 26, 2025
aae7baa
Smollm targets
rohansjoshi Aug 26, 2025
01ca904
Override unload_method in training_module to erase the tensors pointi…
silverguo Aug 26, 2025
4df836d
Disable mm + add -> addmm fusion if added tensor rank >2
hsharma35 Aug 26, 2025
9d6a7f2
Fix bad optimized kernel for add.
hsharma35 Aug 26, 2025
e1cd63e
Allow zero-element inputs for method.
hsharma35 Aug 26, 2025
f93d524
Arm backend: Use cmake for building in Ethos-U jupyter example (#13630)
AdrianLundell Aug 26, 2025
36cdaec
NXP backend: Add MobileNetV2 example model and test (#12892)
StrycekSimon Aug 26, 2025
7bb115b
Arm Backend: Add support for ELU.default operator
agrima1304 Jul 15, 2025
c9cbad7
Arm Backend: Add support for ELU.default operator
agrima1304 Jul 15, 2025
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
The table of contents is too big for display.
Diff view
Diff view
  •  
  •  
  •  
Empty file.
Empty file.
2 changes: 1 addition & 1 deletion .ci/docker/ci_commit_pins/optimum-executorch.txt
Original file line number Diff line number Diff line change
@@ -1 +1 @@
eea657ddbdeb1118943a92fb73c289985c3ee1ba
36e3dd54effb3f6d13d792029609292fdd5502bb
2 changes: 1 addition & 1 deletion .ci/docker/ci_commit_pins/pytorch.txt
Original file line number Diff line number Diff line change
@@ -1 +1 @@
6fc0ad22f0a07b6f38d138861c56a765d5a9bb02
e7152ff8a6a929a0db7f3f4a72a5b6d471769cd3
2 changes: 1 addition & 1 deletion .ci/scripts/build-mediatek-sdk.sh
Original file line number Diff line number Diff line change
Expand Up @@ -14,9 +14,9 @@ build_neuron_backend() {
export NEURON_BUFFER_ALLOCATOR_LIB=${MEDIATEK_SDK_ROOT}/libneuron_buffer_allocator.so
export EXECUTORCH_ROOT="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")/../.." && pwd)"


cd ${EXECUTORCH_ROOT}
./backends/mediatek/scripts/mtk_build.sh
./examples/mediatek/mtk_build_examples.sh
}

build_neuron_backend
2 changes: 1 addition & 1 deletion .ci/scripts/setup-conda.sh
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ set -ex

install_conda() {
pushd .ci/docker || return
${CONDA_INSTALL} -y --file conda-env-ci.txt
${CONDA_INSTALL} -c conda-forge -y --file conda-env-ci.txt
popd || return
}

Expand Down
1 change: 1 addition & 0 deletions .ci/scripts/setup-linux.sh
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@ set -exu
source "$(dirname "${BASH_SOURCE[0]}")/utils.sh"

read -r BUILD_TOOL BUILD_MODE EDITABLE < <(parse_args "$@")
echo "Build tool: $BUILD_TOOL, Mode: $BUILD_MODE"

# As Linux job is running inside a Docker container, all of its dependencies
# have already been installed, so we use PyTorch build from source here instead
Expand Down
1 change: 1 addition & 0 deletions .ci/scripts/setup-vulkan-linux-deps.sh
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@ install_swiftshader() {

export VK_ICD_FILENAMES="${_swiftshader_dir}/swiftshader/build/Linux/vk_swiftshader_icd.json"
export LD_LIBRARY_PATH="${_swiftshader_dir}/swiftshader/build/Linux/"
export ETVK_USING_SWIFTSHADER=1
}

install_vulkan_sdk() {
Expand Down
57 changes: 57 additions & 0 deletions .ci/scripts/test_backend_linux.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
#!/usr/bin/env bash
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the BSD-style license found in the
# LICENSE file in the root directory of this source tree.
set -eux

SUITE=$1
FLOW=$2
ARTIFACT_DIR=$3

REPORT_FILE="$ARTIFACT_DIR/test-report-$FLOW-$SUITE.csv"

echo "Running backend test job for suite $SUITE, flow $FLOW."
echo "Saving job artifacts to $ARTIFACT_DIR."

# The generic Linux job chooses to use base env, not the one setup by the image
eval "$(conda shell.bash hook)"
CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
conda activate "${CONDA_ENV}"

export PYTHON_EXECUTABLE=python

# CMake options to use, in addition to the defaults.
EXTRA_BUILD_ARGS=""

if [[ "$FLOW" == *qnn* ]]; then
# Setup QNN sdk and deps - note that this is a bit hacky due to the nature of the
# Qualcomm build. TODO (gjcomer) Clean this up once the QNN pybinding integration is
# cleaned up.
PYTHON_EXECUTABLE=python bash .ci/scripts/setup-linux.sh --build-tool cmake
PYTHON_EXECUTABLE=python bash .ci/scripts/setup-qnn-deps.sh
PYTHON_EXECUTABLE=python bash .ci/scripts/build-qnn-sdk.sh
QNN_X86_LIB_DIR=`realpath build-x86/lib/`
QNN_SDK_ROOT="/tmp/qnn/2.28.0.241029"
export LD_LIBRARY_PATH"=$QNN_X86_LIB_DIR:$QNN_SDK_ROOT/lib/x86_64-linux-clang/:${LD_LIBRARY_PATH:-}"

# TODO Get SDK root from install scripts
EXTRA_BUILD_ARGS+=" -DEXECUTORCH_BUILD_QNN=ON -DQNN_SDK_ROOT=$QNN_SDK_ROOT"
fi

if [[ "$FLOW" == *vulkan* ]]; then
# Setup swiftshader and Vulkan SDK which are required to build the Vulkan delegate
source .ci/scripts/setup-vulkan-linux-deps.sh

EXTRA_BUILD_ARGS+=" -DEXECUTORCH_BUILD_VULKAN=ON"
fi

# We need the runner to test the built library.
PYTHON_EXECUTABLE=python CMAKE_ARGS="$EXTRA_BUILD_ARGS" .ci/scripts/setup-linux.sh --build-tool cmake --build-mode Release --editable true

EXIT_CODE=0
python -m executorch.backends.test.suite.runner $SUITE --flow $FLOW --report "$REPORT_FILE" || EXIT_CODE=$?

# Generate markdown summary.
python -m executorch.backends.test.suite.generate_markdown_summary "$REPORT_FILE" > ${GITHUB_STEP_SUMMARY:-"step_summary.md"} --exit-code $EXIT_CODE
30 changes: 30 additions & 0 deletions .ci/scripts/test_backend_macos.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@
#!/usr/bin/env bash
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the BSD-style license found in the
# LICENSE file in the root directory of this source tree.
set -eux

SUITE=$1
FLOW=$2
ARTIFACT_DIR=$3

REPORT_FILE="$ARTIFACT_DIR/test-report-$FLOW-$SUITE.csv"

echo "Running backend test job for suite $SUITE, flow $FLOW."
echo "Saving job artifacts to $ARTIFACT_DIR."

${CONDA_RUN} --no-capture-output pip install awscli==1.37.21

bash .ci/scripts/setup-conda.sh
eval "$(conda shell.bash hook)"

PYTHON_EXECUTABLE=python
${CONDA_RUN} --no-capture-output .ci/scripts/setup-macos.sh --build-tool cmake --build-mode Release

EXIT_CODE=0
${CONDA_RUN} --no-capture-output python -m executorch.backends.test.suite.runner $SUITE --flow $FLOW --report "$REPORT_FILE" || EXIT_CODE=$?

# Generate markdown summary.
${CONDA_RUN} --no-capture-output python -m executorch.backends.test.suite.generate_markdown_summary "$REPORT_FILE" > ${GITHUB_STEP_SUMMARY:-"step_summary.md"} --exit-code $EXIT_CODE
Loading
Loading