116 Pull requests merged by 42 people

• [ET-VK] Add 2D Reduction to Vulkan Backend

  #12860 merged Aug 6, 2025

• -Gate lazy registration logic behind compile time macros

  #13134 merged Aug 6, 2025

• Add palletization/codebook support to CoreML backend

  #13051 merged Aug 6, 2025

• Correct the Windows cross-compiling commands

  #13071 merged Aug 6, 2025

• Added readme for Wasm extension

  #13131 merged Aug 6, 2025

• Cleaned up Wasm cmake

  #13106 merged Aug 6, 2025

• [ET-VK][ez] Make get_tensor() API protected

  #13168 merged Aug 6, 2025

• [ET-VK][BE] Remove usage of vTensorPtr and get_tensor

  #13167 merged Aug 6, 2025

• [ET-VK][ez] Make get_tensor() API protected

  #13157 merged Aug 6, 2025

• [ET-VK][BE] Remove usage of vTensorPtr and get_tensor

  #13156 merged Aug 6, 2025

• migrate etrecord test from exir.capture to torch.export

  #13162 merged Aug 6, 2025

• Qualcomm AI Engine Direct - GA Static Qwen3

  #13086 merged Aug 6, 2025

• Enable conv_former model in CI

  #12930 merged Aug 6, 2025

• Update the oss runner to comptiable the passed tokenizer

  #13160 merged Aug 6, 2025

• migrate etrecord generation after to_edge_transform_and_lower to new infra

  #13153 merged Aug 6, 2025

• Add ability to pass model_dir to .ci/scripts/test_huggingface_optimum

  #13116 merged Aug 6, 2025

• migrate etrecord test from exir.capture to torch.export

  #13148 merged Aug 6, 2025

• Use temp allocator for kernel registry

  #13012 merged Aug 6, 2025

• Update remove clone to drop no-op q/dq

  #10920 merged Aug 6, 2025

• NXP backend: Use default (non-shared) quantization params for HardTanh

  #12893 merged Aug 6, 2025

• NXP backend: Fix Copyright findings

  #13125 merged Aug 6, 2025

• migrate etrecord generation after to_edge_transform_and_lower to new infra

  #13058 merged Aug 6, 2025

• Arm backend: Added in VGF unit tests for models

  #13102 merged Aug 6, 2025

• Arm backend: Fix bug in compile spec for VGF

  #13059 merged Aug 6, 2025

• Arm backend: Add tests for TOSA and VGF for extract_io_params_tosa + fix for bug with is_U55_subset

  #13065 merged Aug 6, 2025

• Enable both aten and exir for ops in program builder.

  #13075 merged Aug 6, 2025

• Add CoreML BERT models to CI

  #13105 merged Aug 6, 2025

• Update using-executorch-ios.md

  #13142 merged Aug 5, 2025

• Update using-executorch-ios.md

  #13029 merged Aug 5, 2025

• Qualcomm AI Engine Direct - Runtime Option

  #12297 merged Aug 5, 2025

• Update tutorial-arm-ethos-u.md

  #12979 merged Aug 5, 2025

• Fix typos in docs

  #12969 merged Aug 5, 2025

• Update using-executorch-ios.md

  #12946 merged Aug 5, 2025

• Update llama README to reflect kleidiai is default

  #12905 merged Aug 5, 2025

• remove buck cache and metadata when ./installaton.sh --clean

  #12830 merged Aug 5, 2025

• Check for buffer overflow in prim_ops::et_copy_index()

  #12827 merged Aug 5, 2025

• Updating Docs for Arm Ethos U Tutorial

  #12823 merged Aug 5, 2025

• forward fix

  #13129 merged Aug 5, 2025

• - Move register backend logic to its own file

  #13014 merged Aug 5, 2025

• fix the binary path

  #12820 merged Aug 5, 2025

• make etrecord set representive IO

  #13130 merged Aug 5, 2025

• make etrecord set representive IO

  #13052 merged Aug 5, 2025

• xplat/executorch/runtime/core/portable_type/c10/c10/targets.bzl

  #13103 merged Aug 5, 2025

• [ET-VK] 7/n Split dispatches between multiple command buffers. Split execute dispatch into multiple commands based on dispatch count.

  #13118 merged Aug 5, 2025

• Arm backend: Move q/dq ops constants to backends/arm/constants.py

  #13095 merged Aug 5, 2025

• Arm backend: Move quant util functions to closer to usage

  #13094 merged Aug 5, 2025

• Extract common helper and constants into sepearate files for reusability

  #13109 merged Aug 5, 2025

• Arm backend: Remove unused functions

  #13093 merged Aug 5, 2025

• Arm backend: Import executorch deps using find_package

  #13060 merged Aug 5, 2025

• bring etrecord updated "reverted" by gh patch fix bot back

  #13117 merged Aug 5, 2025

• Readme say set ANDROID_SDK

  #13112 merged Aug 5, 2025

• Run Qwen script with Buck

  #13114 merged Aug 5, 2025

• [ET-VK] 7/n Split dispatches between multiple command buffers. Split execute dispatch into multiple commands based on dispatch count.

  #12530 merged Aug 4, 2025

• Static attention: support local-global attention

  #13043 merged Aug 4, 2025

• Add ISSRuntimeFailure to improve exception handling

  #13069 merged Aug 4, 2025

• Cortex-m Size bump by 1.5KiB

  #13064 merged Aug 4, 2025

• Re-order the model list in CI alphabetical order

  #12929 merged Aug 4, 2025

• Change default minimum_deployment_target to None and support CoreML models with no inputs

  #13053 merged Aug 4, 2025

• Update checksums in Package.swift

  #13108 merged Aug 4, 2025

• Qualcomm AI Engine Direct - Fix the regression of whisper model

  #13062 merged Aug 4, 2025

• Qualcomm AI Engine Direct - Support simple_eval in calibration, perpl…

  #12958 merged Aug 4, 2025

• Use ValueConstructible for execute/forward return type in Module.

  #13090 merged Aug 4, 2025

• Add CoreML CI jobs for optimum-executorch models

  #12870 merged Aug 4, 2025

• NXP backend: Add model input and output quantization

  #12586 merged Aug 4, 2025

• Arm backend: Add arm-baremetal CMake preset

  #13042 merged Aug 4, 2025

• Arm backend: Add VGF unit tests to operators (Part 2)

  #13033 merged Aug 4, 2025

• Arm backend: Add VGF unit tests to operators (Part 1)

  #13032 merged Aug 4, 2025

• NXP backend: Support for Sigmoid operator conversion

  #13041 merged Aug 4, 2025

• NXP backend: Make the flow robust against input/output swapping.

  #12890 merged Aug 4, 2025

• Introduce ValueConstructible protocol.

  #13089 merged Aug 4, 2025

• Use variadic arguments for execute/forward instead of a single one.

  #13088 merged Aug 4, 2025

• Make Tensor data accessors non-throwing.

  #13087 merged Aug 4, 2025

• Refactor MTK build scripts Test building MTK examples and runners in ci

  #13085 merged Aug 4, 2025

• Extend op_add for complex dtype

  #12977 merged Aug 2, 2025

• make to_edge_transform_and_lower support etrecord generation

  #13034 merged Aug 2, 2025

• add executorch program equipment support in etrecord class

  #13020 merged Aug 2, 2025

• add more export modules after ertrecod created

  #13010 merged Aug 2, 2025

• Remove support for VkSemaphore

  #13070 merged Aug 1, 2025

• Fix typo in call to gen_selected_ops

  #13072 merged Aug 1, 2025

• Fix installation that causes torch conflict

  #13074 merged Aug 1, 2025

• Decouple LlavaRunner from multimodal runner

  #13067 merged Aug 1, 2025

• Introduce the get_fake_quant_model API

  #12997 merged Aug 1, 2025

• Pin bump Cpuinfo to 33ed0be

  #13025 merged Aug 1, 2025

• Call ExportPass() inside ReplaceNopTransposeOrPermuteWithViewPass::call().

  #13005 merged Aug 1, 2025

• CI test for lora

  #13068 merged Aug 1, 2025

• [ET-VK] 6/n Split dispatches between multiple command buffers. Replaced encode_execute function with invalidate_execute_encoding and moved encoding logic to execute function().

  #13054 merged Aug 1, 2025

• CI test for lora

  #12724 merged Aug 1, 2025

• Fix MTK build scripts

  #13061 merged Aug 1, 2025

• Add Half / Bfloat16 Tests

  #13048 merged Aug 1, 2025

• [ExecuTorch][Export][1/N] Export API pipeline re-architecture, making it composable

  #13055 merged Aug 1, 2025

• [Executorch][Export][3/N] Modularize export recipes

  #13057 merged Aug 1, 2025

• [Executorch][Export][2/N] Add to_edge and to_backend stages

  #13056 merged Aug 1, 2025

• [Executorch][Export][3/N] Modularize export recipes

  #12938 merged Aug 1, 2025

• [Executorch][Export][2/N] Add to_edge and to_backend stages

  #12937 merged Aug 1, 2025

• [ExecuTorch][Export][1/N] Export API pipeline re-architecture, making it composable

  #12936 merged Aug 1, 2025

• [ET-VK] 6/n Split dispatches between multiple command buffers. Replaced encode_execute function with invalidate_execute_encoding and moved encoding logic to execute function().

  #13016 merged Aug 1, 2025

• Do not use BNNS copy when dtypes differ in CoreML

  #13018 merged Aug 1, 2025

• Mark Tensor and Value final.

  #13049 merged Jul 31, 2025

• [Android] Move training part to its own package

  #13047 merged Jul 31, 2025

• Allow overriding CoreML op support in partitioner to ignore ops where CoreML has bugs

  #13023 merged Jul 31, 2025

• Arm backend: Migrate asinh tests to TOSA 1.0

  #13045 merged Jul 31, 2025

• Arm backend: Create ethosu directory (#11849)

  #12338 merged Jul 31, 2025

• Add exception handling for numerical errors

  #13026 merged Jul 31, 2025

• Extend cat op for complex dtype

  #12894 merged Jul 31, 2025

• Arm backend: Extract code to helpers in executor runner

  #13039 merged Jul 31, 2025

• Arm backend: Seperate --bundleio and --etdump in run.sh

  #13036 merged Jul 31, 2025

• Arm backend: Remove TOSA v0.80 from arm backend

  #13040 merged Jul 31, 2025

• Arm backend: support list-of-tensors args in FuseConstantArgsPass

  #13037 merged Jul 31, 2025

• Arm backend: Add asinh decomposition pass and test

  #13035 merged Jul 31, 2025

• Arm backend: Eliminate one memory copy in executor runner

  #12992 merged Jul 31, 2025

• Arm backend: Move corstone related build logic to util script

  #12993 merged Jul 31, 2025

• Rewrite Memory Metadata Tagging Pass

  #12927 merged Jul 31, 2025

• Update using-executorch-ios.md

  #13024 merged Jul 31, 2025

• Add unload method to module

  #12984 merged Jul 31, 2025

• equip etrecord class with save method

  #13027 merged Jul 31, 2025

• equip etrecord class with save method

  #12978 merged Jul 31, 2025

48 Pull requests opened by 24 people

• Qualcomm AI Engine Direct - fix typo

  #13030 opened Jul 31, 2025

• Enable transformers multimodal export flow in extension/llm

  #13031 opened Jul 31, 2025

• Expose start_pos in generate_from_pos

  #13046 opened Jul 31, 2025

• streamlined optimum module

  #13050 opened Jul 31, 2025

• Remove unused sleef.h which breaks cross-compilation on windows

  #13079 opened Aug 2, 2025

• Correct the flatc/flatcc compilation for x86 ios simulator

  #13080 opened Aug 2, 2025

• Improve softmax perf when transpose is not needed

  #13081 opened Aug 3, 2025

• [ET-VK][BE] Remove usage of `vTensorPtr` and `get_tensor`

  #13096 opened Aug 4, 2025

• [ET-VK][ez] Make `get_tensor()` API protected

  #13097 opened Aug 4, 2025

• [ET-VK][BE] Move all ops to use `DynamicDispatchNode`

  #13098 opened Aug 4, 2025

• Fix torchao deps

  #13107 opened Aug 4, 2025

• Update install_requirements.py to include TORCH_NIGHTLY_URL for requirement-examples.txt

  #13110 opened Aug 4, 2025

• Enable dinov2 model in CI

  #13111 opened Aug 4, 2025

• [ET-VK] 8/n Split dispatches between multiple command buffers. This diff adds a config to limit the maximum number of command buffers created when splitting execution between multiple command buffers.

  #13113 opened Aug 4, 2025

• Updates to Zephyr CI Models Test

  #13115 opened Aug 4, 2025

• [ET-VK][ez] Small optimization to `DynamicDispatchNode`

  #13119 opened Aug 4, 2025

• Update CoreML docs

  #13120 opened Aug 5, 2025

• [Executorch][Recipes][Coreml] Add coreml backend recipes

  #13121 opened Aug 5, 2025

• Add AOTI backend skeleton code

  #13123 opened Aug 5, 2025

• Arm backend: Move QuantArgs to its own file in _passes/

  #13124 opened Aug 5, 2025

• NXP backend: Node extraction exception enhancement

  #13126 opened Aug 5, 2025

• Qualcomm AI Engine Direct - Static Decoder Runner Support 16bit KV IO

  #13127 opened Aug 5, 2025

• NXP backend: Decrease min-num-ops-per-graph

  #13128 opened Aug 5, 2025

• Refactor op_cat to use util pattern

  #13132 opened Aug 5, 2025

• Test xnnpack with pybindings

  #13133 opened Aug 5, 2025

• [ET-VK][BE] Remove usage of `vTensorPtr` and `get_tensor`

  #13136 opened Aug 5, 2025

• [ET-VK][ez] Make `get_tensor()` API protected

  #13137 opened Aug 5, 2025

• [ET-VK][BE] Move all ops to use `DynamicDispatchNode`

  #13138 opened Aug 5, 2025

• [ET-VK][ez] Small optimization to `DynamicDispatchNode`

  #13139 opened Aug 5, 2025

• Extend tranform() method params with PassManager.

  #13140 opened Aug 5, 2025

• raise error when trying to save an etrecord missing essential info

  #13143 opened Aug 5, 2025

• Fix et_logger by remove the self args in the et_logger decorator

  #13144 opened Aug 5, 2025

• Refactor op_stack to use util pattern

  #13146 opened Aug 6, 2025

• enable qat for custom annotation in qnn

  #13147 opened Aug 6, 2025

• make ExecuTorch program always has valid debug handle map and delegate map

  #13149 opened Aug 6, 2025

• Qualcomm AI Engine Direct - Refactor calibration flow

  #13150 opened Aug 6, 2025

• make to_edge support etrecord generation

  #13151 opened Aug 6, 2025

• Fix the stride issue in DecomposeAvgPool2d and add a test for it.

  #13152 opened Aug 6, 2025

• Arm backend: Add VGF unit tests to operators (Part 3)

  #13154 opened Aug 6, 2025

• Arm Backend: Improve VGF runtime and update MLSDK setup

  #13155 opened Aug 6, 2025

• [ET-VK][BE] Move all ops to use `DynamicDispatchNode`

  #13158 opened Aug 6, 2025

• [ET-VK][ez] Small optimization to `DynamicDispatchNode`

  #13159 opened Aug 6, 2025

• Save foundation weights separately

  #13161 opened Aug 6, 2025

• Use unlifted export pass to tag delegated constants

  #13163 opened Aug 6, 2025

• make et.export support etrecord generation

  #13164 opened Aug 6, 2025

• Arm backend: run on tosa_ref_model only when requested

  #13165 opened Aug 6, 2025

• Add a generic multimodal runner

  #13166 opened Aug 6, 2025

• fix llama buck build

  #13169 opened Aug 6, 2025

24 Issues closed by 8 people

• [Android] Move SGD and TrainingModule org.pytorch.executorch.extension.training ?

  #13028 closed Aug 5, 2025

• CoreML fails to lower addmm with integer alpha/beta parameters

  #11689 closed Aug 4, 2025

• exir "missing out vars"

  #3443 closed Aug 4, 2025

• Error when build coreml_executor_runner

  #5020 closed Aug 4, 2025

• XNN int8 out of bounds write / invalid outputs

  #10960 closed Aug 4, 2025

• Run CoreML backend tests in CI

  #9115 closed Aug 4, 2025

• dl3 model fails to export with coreml delegates

  #5159 closed Aug 4, 2025

• Llama 3.2 3B Core ML poor output quality

  #9395 closed Aug 4, 2025

• CoreML ignores add/sub alpha parameter

  #11687 closed Aug 4, 2025

• CoreML gives wrong outputs or crashes for torch.diagonal

  #11718 closed Aug 4, 2025

• CoreML floor_divide crashes process

  #11714 closed Aug 4, 2025

• CoreML model with no inputs fails to load at runtime

  #11719 closed Aug 4, 2025

• CoreML randint occasionally generates out of distribution values

  #11721 closed Aug 4, 2025

• CoreML fails to lower asinh/acosh

  #11712 closed Aug 4, 2025

• Modularize export recipes

  #12933 closed Aug 4, 2025

• Add support for to_edge and to_backend stages

  #12932 closed Aug 4, 2025

• Make et/export pipeline composable

  #12928 closed Aug 4, 2025

• Failed build of apple-ios example

  #11753 closed Aug 2, 2025

• Problems with Swift/iOS

  #13066 closed Aug 1, 2025

• Xnnpack export fails with negative dim in ConstantPad2d

  #12456 closed Jul 31, 2025

• [Delegate Testing] Expand to individual op level tests for core-aten, quantized and non-quantized

  #12898 closed Jul 31, 2025

• Update LLM documents and websites

  #12214 closed Jul 31, 2025

• Doc fix for Android Llama Demo tutorial

  #12767 closed Jul 31, 2025

• Build from source doesn't work

  #12762 closed Jul 31, 2025

16 Issues opened by 9 people

• Question about the code in executorch/examples/qualcomm/oss_scripts/llama/llama.py

  #13145 opened Aug 6, 2025

• TOSA v0.80 still presents in nightly dev build after removal PR

  #13141 opened Aug 5, 2025

• Qwen3 Vulkan Delegate Failures at Runtime

  #13135 opened Aug 5, 2025

• quantize_ in CoreML only works on iOS18 and the error message is confusing

  #13122 opened Aug 5, 2025

• Support GraphModule and ExportedProgram including nn.module

  #13104 opened Aug 4, 2025

• Add QNN backend recipes

  #13101 opened Aug 4, 2025

• Add CoreML backend recipes

  #13100 opened Aug 4, 2025

• Support QAT Quantization in executorch export

  #13099 opened Aug 4, 2025

• QNN backend. The fake quant model (the graph after convert_pt2e) does not generate accurate result

  #13092 opened Aug 4, 2025

• QNN, The performance gap between nchw and ncwh is huge. When either h or w is 1,

  #13091 opened Aug 4, 2025

• [Vulkan] Make partitioning dtype aware

  #13084 opened Aug 4, 2025

• [Vulkan] Fix 64-bit I/O in Vulkan Delegate

  #13083 opened Aug 4, 2025

• [Vulkan] Add (limited) support for >4-dim tensors

  #13082 opened Aug 4, 2025

• NXP Backend: Quantize selected activation functions with Conv/Matrix multiplication

  #13063 opened Aug 1, 2025

• Publish pypi wheel for linux aarch64

  #13044 opened Jul 31, 2025

• Incorrect display of logs for ARM bulds on FVP

  #13038 opened Jul 31, 2025

59 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Add XNNPACK backend option for workspace sharing

  #11748 commented on Aug 4, 2025 • 30 new comments

• Add _clone_dim_order portable kernel

  #12974 commented on Aug 6, 2025 • 10 new comments

• Arm backend: Make TOSA backend NCHW-compatible

  #12994 commented on Aug 6, 2025 • 8 new comments

• Arm Zephyr CI Test for Whisper-tiny Encoder and Decoder

  #13008 commented on Aug 1, 2025 • 6 new comments

• Arm Backend: Add support for ELU.default operator

  #12996 commented on Aug 1, 2025 • 6 new comments

• Add initial backends/cadence/vision module scaffold with optimized softmax kernel (no iDMA)

  #12480 commented on Aug 4, 2025 • 2 new comments

• NXP backend: Add MobileNetV2 example model and test

  #12892 commented on Aug 5, 2025 • 1 new comment

• Arm backend: Add dump_delegate_data function

  #12334 commented on Jul 31, 2025 • 0 new comments

• Qualcomm AI Engine Direct - GA Qwen 2.5 0.5B

  #12333 commented on Aug 6, 2025 • 0 new comments

• Arm backend: Add Inception v3 test

  #12111 commented on Aug 6, 2025 • 0 new comments

• Arm backend: Add VIT test

  #11996 commented on Jul 31, 2025 • 0 new comments

• Arm backend: Replace Tensor_Scalar with Tensor_Tensor in sqrt op

  #11783 commented on Jul 31, 2025 • 0 new comments

• Only support int8 and quant dtypes for quant operators

  #11685 commented on Aug 6, 2025 • 0 new comments

• Android preset

  #11119 commented on Aug 6, 2025 • 0 new comments

• Android build cmake in gradle

  #11085 commented on Aug 5, 2025 • 0 new comments

• [pytorch hash update] update the pinned pytorch hash

  #10955 commented on Aug 6, 2025 • 0 new comments

• [devtool] create stream_data_sink

  #8604 commented on Aug 6, 2025 • 0 new comments

• [v0.7.0] Release Tracker

  #11075 commented on Jul 31, 2025 • 0 new comments

• PAL File for Arm BareMetal Executor Runner

  #12399 commented on Aug 6, 2025 • 0 new comments

• Cmsis

  #12655 commented on Jul 31, 2025 • 0 new comments

• Fix temp memory allocation issue in torch.topk operations

  #12810 commented on Aug 4, 2025 • 0 new comments

• Extend `PyBundledModule` with `extension.BundledModule`

  #12839 commented on Aug 5, 2025 • 0 new comments

• Support for Dtye Selective Build from Dictionary API in OSS (cmake)

  #12873 commented on Aug 1, 2025 • 0 new comments

• [EXIR] Register _clone_dim_order op and map aten.clone

  #12971 commented on Aug 6, 2025 • 0 new comments

• [EXIR] Update RemoveCloneOpsTransform to be dim order aware

  #12976 commented on Aug 6, 2025 • 0 new comments

• WIP - Executorch Theme Migration

  #12989 commented on Aug 5, 2025 • 0 new comments

• Arm backend: Add full partition rejections

  #12995 commented on Aug 6, 2025 • 0 new comments

• Move to Span<EValue*> instead of EValue** in delegate interface

  #13004 commented on Aug 6, 2025 • 0 new comments

• Default Export for Mv2 with Real Image Input

  #13019 commented on Jul 31, 2025 • 0 new comments

• Arm backend: Add a simple weight sharing test

  #13021 commented on Jul 31, 2025 • 0 new comments

• Share int64 → int32 cast insertion pass across backends (TOSA, XNNPACK, etc.)

  #12957 commented on Jul 31, 2025 • 0 new comments

• Android native crash when running DeepLabV3Demo with vulkan backend in Android 16

  #11754 commented on Jul 31, 2025 • 0 new comments

• make export flow etrecord generation supportive

  #12961 commented on Jul 31, 2025 • 0 new comments

• [Cortex-M] Infer Quantized MV2 or 3 performance on par with TFLM - with CMSIS-NN

  #12492 commented on Jul 31, 2025 • 0 new comments

• [Android] Add method metadata report API in Module.java

  #10453 commented on Jul 31, 2025 • 0 new comments

• [CoreML] Enable palletization via quantize_

  #12923 commented on Jul 31, 2025 • 0 new comments

• Add heap based greedy planning algorithm and graph coloring algorithm for memory planning.

  #8467 commented on Aug 3, 2025 • 0 new comments

• Fails to import EthosUPartitioner for ARM backend as there is numpy inconsistency with tosa-tools

  #12991 commented on Aug 4, 2025 • 0 new comments

• Fix typo ${tolchain} in build_portable_kernels.sh

  #12858 commented on Aug 4, 2025 • 0 new comments

• Fix typo ${tolchain} in build_executor_runner.sh

  #12859 commented on Aug 4, 2025 • 0 new comments

• CoreML argmin/argmax intermittently crashes process

  #11715 commented on Aug 4, 2025 • 0 new comments

• QNN parallel performance issues

  #12990 commented on Aug 4, 2025 • 0 new comments

• Running Qwen3 in Android LlamaDemo App shows an error while loading tokenizer

  #11311 commented on Aug 4, 2025 • 0 new comments

• --skip-fvp-setup flag is ignored

  #12447 commented on Aug 5, 2025 • 0 new comments

• Freeing memory using free() which was not allocated by malloc()

  #11395 commented on Aug 5, 2025 • 0 new comments

• Arm backend: Vela internal error

  #13022 commented on Aug 5, 2025 • 0 new comments

• NXP backend: Update atol in `test_hardtanh_converter.py` tests

  #12951 commented on Aug 5, 2025 • 0 new comments

• Support for Shared Weights between multiple entry points in PTE export

  #12804 commented on Aug 5, 2025 • 0 new comments

• NXP backend: Move calibration inputs creation outside to_quantized_edge_program

  #12950 commented on Aug 5, 2025 • 0 new comments

• Lack of AVX-512 causing installation errors while installing Executorch

  #12817 commented on Aug 6, 2025 • 0 new comments

• Incorrect Observer Sharing/Derivation at Conv-ReLU+ Residual with Arm Ethos Quantizer

  #12959 commented on Aug 6, 2025 • 0 new comments

• [RFC][Short] More on ExportRecipes

  #12660 commented on Aug 6, 2025 • 0 new comments

• efficient_sam fails to run on CoreML

  #12906 commented on Aug 6, 2025 • 0 new comments

• fix spec_prop_pass

  #7974 commented on Aug 6, 2025 • 0 new comments

• Fix comment in memory_planning.py

  #8010 commented on Aug 6, 2025 • 0 new comments

• remove the exec_aten namespace

  #8018 commented on Aug 6, 2025 • 0 new comments

• Revert to use mean_out than mean_dim_out

  #8021 commented on Aug 6, 2025 • 0 new comments

• [docs][ez] Fix doc build workflow

  #8079 commented on Aug 6, 2025 • 0 new comments

• Adjust tolerance for quantized XNN conv1d tests

  #8093 commented on Aug 6, 2025 • 0 new comments