disclaimer #3376

mergennachin · 2024-04-26T18:28:10Z

Use unified path /data/local/tmp/llama ([Android] Use unified path /data/local/tmp/llama #2899)
Update docs for the demo app. (Update docs for the demo app. #2921)
Update docs (Update docs #2919)
exclude mutated buffer (exclude mutated buffer #2876)
Update iphone 15 pro benchmarking numbers (Update iphone 15 pro benchmarking numbers #2927)
Fixing minor issues in llama2 7b repro (Fixing minor issues in llama2 7b repro #2926)
Refine LLM getting started guide for uniformity, fix critical errors (Refine LLM getting started guide for uniformity, fix critical errors #2911)
Prefer native buildifier suppressions over @lint-ignore BUILDIFIERLINT
Fix generation speed calculation. (Fix generation speed calculation. #2932)
Qualcomm AI Engine Direct - HardSigmoid follow up for FP16 / Test cases complement (Qualcomm AI Engine Direct - HardSigmoid follow up for FP16 / Test cases complement #2790)
Move legacy range constraint calculator to executorch to unblock pytorch CI (Move legacy range constraint calculator to executorch to unblock pytorch CI #2925)
Bump version 0.2.0->0.3.0 ([release] Bump version 0.2.0->0.3.0 #2703)
Rename to max_pool2d.yaml ([ET-VK][EZ] Rename to max_pool2d.yaml #2820)
Assign trivially moveable values to const-ref variables ([ET-VK][EZ] Assign trivially moveable values to const-ref variables #2834)
Introduce add_tensor_like consuming ValueRef ([ET-VK] Introduce add_tensor overloads consuming TensorRef #2835)
Refactor Pool.cpp ([ET-VK] Refactor Pool.cpp #2836)
Fix comment style for parameter names ([ET-VK][EZ] Fix comment style for parameter names #2847)
aten.convolution (SlidingWindow) ([ET-VK][Ops] aten.convolution (SlidingWindow) #2812)
Add executorch_no_prim_ops target (Add executorch_no_prim_ops target #2934)
oss: Add buck_build cfg (oss: Add buck_build cfg #2902)
update torch pin (update torch pin #2944)
aten.convolution (Transpose) ([ET-VK][Ops] aten.convolution (Transpose) #2883)
aten.convolution (Depthwise) ([ET-VK][Ops] aten.convolution (Depthwise) #2884)
Fix Validation Layer warnings about wrong image layout ([ET-VK][Ez] Fix Validation Layer warnings about wrong image layout #2854)
Introduce convenience constexpr for StorageTypes and GPUMemoryLayouts ([ET-VK][Ez] Introduce convenience constexpr for StorageTypes and GPUMemoryLayouts #2948)
Use __ET_UNLIKELY in assertion macros ([executorch] Use __ET_UNLIKELY in assertion macros #2949)
s/heirarchies/hierarchies/ (s/heirarchies/hierarchies/ #2772)
Fix indentation in selective build example code (Fix indentation in selective build example code #2773)
aten.convolution (Depthwise Output-Tile) ([ET-VK][Ops] aten.convolution (Depthwise Output-Tile) #2885)
aten.convolution (Pointwise) ([ET-VK][Ops] aten.convolution (Pointwise) #2886)
Make minor updates to LLM guide setup instructions (Make minor updates to LLM guide setup instructions #2940)
resolve_buck.py: Add an entry for darwin-x86_64 (resolve_buck.py: Add an entry for darwin-x86_64 #2868)
Compute graph print readable (Compute graph print readable #2825)
aten.convolution (Bias=False) ([ET-VK][Ops] aten.convolution (Bias=False) #2887)
Add convolution cases to codegen ([ET-VK] Add convolution cases to codegen #2920)
add aten.sum.default (add aten.sum.default #2807)
Fix failing CI jobs caused by Add executorch_no_prim_ops target #2934 (Fix failing CI jobs caused by #2934 #2961)
Replace std::stringstream with std::string for Shader names ([ET-VK][ez] Replace std::stringstream with std::string for Shader names #2964)
Refine the LLM manual (focus on the debugging and profiling part) (Refine the LLM manual (focus on the debugging and profiling part) #2952)
Android demo app tutorial fix for XNNPACK and QNN (Android demo app tutorial fix for XNNPACK and QNN #2962)
Qualcomm AI Engine Direct - Enable per channel linear op (Qualcomm AI Engine Direct - Enable per channel linear op #2822)
Custom ops API small fixes (Custom ops API small fixes #2936)
Consolidate EXECUTORCH_BUILD_CUSTOM option (Consolidate EXECUTORCH_BUILD_CUSTOM option #2935)
Consolidate tokenizer interface (Consolidate tokenizer interface #2954)
Update OSS repo (Update OSS repo #2033)
Add the missing import generate_etrecord to doc Getting Started with LLM (Getting started with LLM missing import generate_etrecord #2977)
Fix llama runner test (Fix llama runner test #2981)
Forward fix macOS job after test-infra Unified Android aar support for llava and llama models #5086 (Forward fix macOS jobs after test-infra #5086 #2980)
Add a mock perf test for llama2 on Android (Add a mock perf test for llama2 on Android #2963)
Core ML Has Added Index_Put Support, No Need to Skip Anymore (Core ML Has Added Index_Put Support, No Need to Skip Anymore #2975)
Minor fix in README.md page
Add llama2 readme in examples/README (Add llama2 readme in examples/README #2992)
Use new API to register custom ops for llama model (Use new API to register custom ops for llama model #2916)
Fix tutorial for Qualcomm AI Engine Direct Backend (Fix tutorial for Qualcomm AI Engine Direct Backend #2956)
Update name from xtensa to cadence (Update name from xtensa to cadence #2982)
Use new API to register custom ExecuTorch kernels into ATen (Use new API to register custom ExecuTorch kernels into ATen #2937)
fix et-view (fix et-view #2843)
Replace view copy with view (3/3) (Replace view copy with view (3/3) #2463)
Skip annotate boolean input (Skip annotate boolean input #2957)
Fix build-framework-ios CI job (Fix build-framework-ios CI job #2996)
Extend constant prop pass to work with int/float/etc scalars and fix input specs. (Extend constant prop pass to work with int/float/etc scalars and fix input specs. #2950)
Introduce vTensorPtr to prevent reference invalidation and remove get_val() API ([ET-VK] Introduce vTensorPtr to prevent reference invalidation and remove get_val() API #2978)
Add Tiktoken in python (Add Tiktoken in python #2986)
Dynamic Shapes (Dynamic Shapes #2442)
dynamic qd8-fc test with 2 batch dims (dynamic qd8-fc test with 2 batch dims #2441)
dynamic mobilenetv2 (dynamic mobilenetv2 #2440)
dynamic mv3 (dynamic mv3 #2475)
Dynamic ResNet (Dynamic ResNet #2474)
Dynamic ViT (Dynamic ViT #2476)
add export configs (add export configs #2965)
Add exir.save and exir.load with export_serialize (Add exir.save and exir.load with export_serialize #3000)
Fix 3 CI jobs (Fix 3 CI jobs #3006)
Add util to print out ops and frequency (Add util to print out ops and frequency #2983)
Decouple custom ops in llama_transformer.py Part 1/N (Decouple custom ops in llama_transformer.py Part 1/N #3005)
Decouple custom ops in llama_transformer.py Part 2/N (Decouple custom ops in llama_transformer.py Part 2/N #3007)
Update README.md (Update README.md #3012)
add more instructions and examples on Delegation (add more instructions and examples on Delegation #2973)
Run LlamaDemo app on AWS Device Farm (Run LlamaDemo app on AWS Device Farm #3004)
Remove RemoveRedundantViewCopyPass (Remove RemoveRedundantViewCopyPass #2464)
Change tokenizer name to bpe_tokenizer and extract a base class (Change tokenizer name to bpe_tokenizer and extract a base class #3009)
Update README.md and add submodule update (Update README.md and add submodule update #3029)
Throw in VK_GET_OP_FN if op is not found ([ET-VK][EZ] Throw in VK_GET_OP_FN if op is not found #3028)
update the pinned pytorch hash ([pytorch hash update] update the pinned pytorch hash #2824)
Apply clang-format 18
oss: Upgrade clap, add string feature (oss: Upgrade clap, add string feature #3035)
Update to clang 18.1.3
Fix handling constant inputs when delegating (Fix handling constant inputs when delegating #3031)
Fix lint in clang-format (: Fix lint in clang-format #3041)
generation.py with kv cache ([executorch] generation.py with kv cache #3030)
Clean up shader library and introduce some new conventions ([ET-VK] Clean up shader library and introduce some new conventions #3024)
Move compile spec to ArmTester interface (Move compile spec to ArmTester interface #2991)
remove duplicate generate_lib_aten target under aten kernel (remove duplicate generate_lib_aten target under aten kernel #2951)
native_layer_norm (for width dim) (native_layer_norm (for width dim) #3001)
aten.full.default (aten.full.default #3013)
Add tiktoken (Add tiktoken #3015)
Add tiktoken to eval ([executorch][llm] Add tiktoken to eval #3044)
Update pytorch commit pin to 04/15 (Update pytorch commit pin to 04/15 #3047)
Dynamic Conv1d + W2L (Dynamic Conv1d + W2L #2976)
Fix iOS build by excluding external CoreML SDK dependencies (Fix iOS build by excluding external CoreML SDK dependencies #3043)
aten.select.int (aten.select.int #3033)
4b quantized embedding table operator (4b quantized embedding table operator #3050)
Fix test_llama_runner by hiding tiktoken (Fix test_llama_runner by hiding tiktoken #3055)
Bump Vulkan API requirement to 1.1 and enable 16 bit and 8 bit types in buffer storage ([ET-VK] Bump Vulkan API requirement to 1.1 and enable 16 bit and 8 bit types in buffer storage #3058)
Enable FP16 type in operators ([ET-VK] Enable FP16 type in operators #3059)
Fix formatting issues in executorch/test/size_test.cpp (Fix formatting issues in executorch/test/size_test.cpp #3065)
ETRecord ser/de handling "None" outputs and more (ETRecord ser/de handling "None" outputs and more #3039)
Update doc-build.yml (Update doc-build.yml #3045)
Add int16 support to aten_bridge (Add int16 support to aten_bridge #3069)
fix linear recomposition (fix linear recomposition #3064)
Set kernel default visibility to hidden (Set kernel default visibility to hidden #3060)
Fix Android llama2 demo app after Android demo app tutorial fix for XNNPACK and QNN #2962 (Fix Android llama2 demo app after #2962 #3032)
Update doc-build.yml (Update doc-build.yml #3071)
Handle empty (size=0) tensor in Inspector (Handle empty (size=0) tensor in Inspector #2998)
Add quantized op support to llama runner (Add quantized op support to llama runner #3062)
{executorch][llama] support mqa ({executorch][llama] support mqa #3080)
Load missing state dict in edge program serialization (Load missing state dict in edge program serialization #3076)
Remove noindex from upload to gh-pages job (Remove noindex from upload to gh-pages job #3077)
forward fix ConstantArgument initialization (forward fix ConstantArgument initialization #3074)
Fix llama2 README.md cmake instructions (Fix llama2 README.md cmake instructions #3096)
Fix build time warning (Fix build time warning #3097)
change call_delegate_autograd (change call_delegate_autograd #3073)
remove exir.capture from dynamic_shape_propogation test (remove exir.capture from dynamic_shape_propogation test #3070)
Create init.py in example folder (Create __init__.py in example folder #3093)
move mask as sdpa input instead of attribute (move mask as sdpa input instead of attribute #3036)
remove exir.capture from test_rpc.py (remove exir.capture from test_rpc.py #3102)
Introduce SpecVarList to represent specialization constants ([ET-VK] Introduce SpecVarList to represent specialization constants #3078)
Enable additional specialization constants in compute shaders ([ET-VK] Enable additional specialization constants in compute shaders #3079)
select_copy.int ([ET-VK][2/n] select_copy.int #3085)
aten.permute_copy.default ([ET-VK][3/n] aten.permute_copy.default #3086)
Improve codegen for aten.permute ([ET-VK][4/n] Improve codegen for aten.permute #3087)
make_seq_tensor in codegen ([ET-VK][5/n] make_seq_tensor in codegen #3088)
remove exir.capture from quant fusion test (remove exir.capture from quant fusion test #3106)
Don't crash when execute_method fails (Don't crash when execute_method fails #3104)
update readme to not use exir.capture (update readme to not use exir.capture #3107)
remove exir.capture from example delegate test (remove exir.capture from example delegate test #3101)
throw Java exception when execution fails (throw Java exception when execution fails #3112)
Handle missing data types. ([CoreML Backend] Handle missing data types. #2984)
Documentation for Vulkan Delegate ([ET-VK] Documentation for Vulkan Delegate #3113)
fix embedding_4bit resize (fix embedding_4bit resize #3118)
Delete llama_quantized lib (Delete llama_quantized lib #3119)
Add quantized cmake option back to fix build-apple-framework (Add quantized cmake option back to fix build-apple-framework #3115)
Fix typo in sub & clean up (Fix typo in sub & clean up #3100)
Free Vulkan delegate segments after compileModel (Free Vulkan delegate segments after loading #3116)
Define embedding_4bit (Define embedding_4bit #3121)
cherry-pick: Add required deps to pyproject.toml (cherry-pick: Add required deps to pyproject.toml #3117)
Preserve modelname (Preserve modelname #3122)
fix llama-runner-linux-android (fix llama-runner-linux-android #3127)
Buck build - fix use_tiktoken config
delete exir/experimental (delete exir/experimental #3109)
4b embedding quantizer (4b embedding quantizer (#3081) #3135)
Update README.md (Update README.md #3094)
Adding Gotchas in README.md (Adding Gotchas in README.md #3138)
Update README.md for llama3 (Update README.md for llama3 #3141)
aten.view_copy ([ET-VK][6/n] aten.view_copy #3129)
Update README.md on the evaluation parameters (Update README.md on the evaluation parameters #3139)
Add reference to the llama2 example for llama3 ([Llama3] Add reference to the llama2 example for llama3 #3142)
Update Llama3 perplexity numbers in README.md (Update Llama3 perplexity numbers in README.md #3145)
add cpu device to run eval on cpu (add cpu device to run eval on cpu #3133)
Add a simple sdpa (Add a simple sdpa #3037)
Fix quantized embedding export logic (Fix quantized embedding export logic #3095)
Comply llama2 runner with gcc 11.4 (Comply llama2 runner with gcc 11.4 #3140)
qnn end to end flow for stories model (qnn end to end flow #3038)
Instructions for Llama3 (Instructions for Llama3 #3154)
Fix embedding_4bit out variant (Fix embedding_4bit out variant #3151)
Add link to llama3 README file (Add link to llama3 README file #3156)
make op_split_with_sizes_copy support dynamic shape (make op_split_with_sizes_copy support dynamic shape #3152)
Call destructor explicitly when move constructing Value ([ET-VK] Call destructor explicitly when move constructing Value #3148)
Clean up api::vTensor class ([ET-VK] Clean up api::vTensor class #3149)
Introduce ParamsBindList to prevent needing to pass shared_ptr to bind parameter UBOs ([ET-VK] Introduce ParamsBindList to prevent needing to pass shared_ptr to bind parameter UBOs #3150)
Rename tokenizer file in Xcode. (Rename tokenizer file in Xcode. #3160)
Adding .model tokenizer to selection (Adding .model tokenizer to selection #3163)
Docs for lower smaller models to mps/coreml/qnn (Docs for lower smaller models to mps/coreml/qnn #3146)
Add missing ops for RNNT predictor (Add missing ops for RNNT predictor #3125)
fix test-demo-android (fix test-demo-android #3168)
Slice, with lots of codegen improvements ([ET-VK][7/n] Slice, with lots of codegen improvements #3171)
add kv cache to eval ([llm] add kv cache to eval #3162)
Update model arg name rope_theta to be consistent with those in llama's website (Update model arg name rope_theta to be consistent with those in llama's website #3147)
conv1d, special case
Qualcomm AI Engine Direct - Enable SSD300_VGG16 (Qualcomm AI Engine Direct - Enable SSD300_VGG16 #3010)
conv1d with bias=False
Switch to a dedicated brach for prebuilt packages. (Switch to a dedicated brach for prebuilt packages. #3184)
Use "latest" as the version for prebuilt frameworks. (Use "main" as the version for prebuilt frameworks. #3161)
Deprecate gpu_sizes_ubo() and extents(); also toggle packing layout via specialization constants ([ET-VK] Deprecate gpu_sizes_ubo() and toggle packing layout via specialization shader #3181)
Specify OSX deployment target for python package. (Specify OSX deployment target for python package. #3193)
Specify OSX deployment target for python package. (Specify OSX deployment target for python package. #3194)
Fix linter. (Fix linter. #3195)
support emit sym value from delegate (support emit sym value from delegate #3103)
Update Xcode project to build tiktoken tokenizer for LLaMA 3. (Update Xcode project to build tiktoken tokenizer for LLaMA 3. #3197)
Add quantized ops to pybindings ([executorch] Add quantized ops to pybindings #3206)
Add memory and vector include in managed_tensor.h (Add memory and vector include in managed_tensor.h #3201)
Refactor export_llama_lib.py
Update setup.sh for tokenizer selection (Update setup.sh for tokenizer selection #3207)
Qualcomm AI Engine Direct - Fixed uint16 tensor and linear op (Qualcomm AI Engine Direct - Fixed uint16 tensor and linear op #3196)
Add a pure python wrapper to pybindings.portable_lib (Add a pure python wrapper fo pybindings.portable_lib #3137)
Remove unused extension/aot_util directory (Remove unused extension/aot_util directory #3216)
Create dependabot rule to upgrade TorchFix version (Create dependabot rule to upgrade TorchFix version #3208)
Bring back extents_ubo() as texture_limits_ubo() ([ET-VK] Bring back extents_ubo() as texture_limits_ubo() #3217)
backout the schema definition change (backout the schema definition change #3213)
Update some SDK docs from MVP (update sdk doc from mvp #3212)
Bump the torch pin (Bump the torch pin #3199)
Fix LLAMA app (Fix LLAMA app #3228)
Fix executor_runner_mps and mpsdelegate linking with pybind ([MPS] Fix executor_runner_mps and mpsdelegate linking with pybind #3222)
Update to transformers 4.38 (Update to transformers 4.38 #3227)
Update TorchNightly to 2024.04.22 (Update TorchNightly to 2024.04.22 #3225)
Support llama3 ([Android] Support llama3 #3232)
strip symbol when linking (strip symbol when linking #3234)
fix typo (fix typo #3235)
Bump torchfix from 0.1.1 to 0.5.0 (Bump torchfix from 0.1.1 to 0.5.0 #3220)
Pin CoreMLTools 7.2 (Pin CoreMLTools 7.2 #3170)
Expand visibility of targets needed for executorch_llama2 kernel (Expand visibility of targets needed for executorch_llama2 kernel #3174)
Support tensors in prim_getters (Support tensors in prim_getters #3203)
Enable doc upload for tags, disable for release branches (Enable doc upload for tags, disable for release branches #3153)
Update Core ML Backend Doc (Update Core ML Backend Doc #3188)
bundled program alpha document (bundled program alpha document #3224)
Fix a small inconsistency on the SDK debugging page (Fix a small inconsistency on the SDK debugging page #3247)
Update tutorial (Update tutorial #3242)
update sdk delegate integration (update sdk delegate integration #3246)
Add iPad support to demo apps. (Add iPad support to demo apps. #3251)
Add more prebuilt artifacts ([Android] Add more prebuilt artifacts #3245)
SDK tutorial doc update (SDK tutorial doc update #3238)
conv1d general case (conv1d general case #3223)
move code under executorch/example (move code under executorch/example #3176)
update XNNPACK/README.md (update XNNPACK/README.md #3236)
Update Profiling Section in XNNPACK Delegate Docs (Update Profiling Section in XNNPACK Delegate Docs #3237)
Add allocate_temp method to KernelRuntimeContext (Add allocate_temp method to KernelRuntimeContext #3209)
Inspector APIs page
fix qnn install link (fix qnn install link #3260)
Add index.Tensor and aten.logical_not ([MPS] Add index.Tensor and aten.logical_not #3221)
Fix broken links on the coreml tutorial page (Fix broken links on the coreml tutorial page #3250)
Fix compilation with gcc-9+ ([BE] Fix compilation on older Linux flavors #3262)
Add delegate time scale converter to Inspector (Add delegate time scale converter to Inspector #3240)
Tie quantization of add operands and result together (Tie quantization of add operands and result together #3091)
Add semihosting to cmake for executor_runner (Add semihosting to cmake for executor_runner #3008)
Capture output of Vela and print on error (Capture output of Vela and print on error #3057)
Fix for TOSA BI clamp ops ([Arm backend] Fix for TOSA BI clamp ops #3092)
delegation debug page (delegate debug page #3254)
update memory planning docs (update memory planning docs #3270)
DynamicShim for dlsym user (DynamicShim for dlsym user #3136)
Unsqueeze ([ET-VK][8/n] Unsqueeze #3172)
clone node ([ET-VK][9/n] clone node #3219)
add dynamic export into llm manual (add dynamic export into llm manual #3202)
Update readme. (Update readme. #3301)
Fix portable is[inf|nan|_out compilation on older Linux (Fix portable is[inf|nan|_out compilation on older Linux #3272)
Use relative links in llm/getting-started.md (Use relative links in llm/getting-started.md #3244)
Update examples/README.md with Llama 3 and names (Update examples/README.md with Llama 3 and names #3275)
Revert D56480274: Add more prebuilt artifacts
update typos (update typos #3300)
Update readme.
Update custom kernel registration API
llama2 readme (llama2 readme #3315)
Fix sdk_example_runner.sh (Fix sdk_example_runner.sh #3298)
Update readme.
Update MPS documentation; add helper script to build mps_executor_runner ([MPS] Update MPS documentation; add helper script to build mps_executor_runner #3324)
Remove the sorting of the nodes from partitioning (not needed for now as Custom Metal kernels are yet not enabled) ([MPS] Remove the sorting of the nodes from partitioning (not needed for now as Custom Metal kernels are yet not enabled) #3328)
copy node, aten.repeat ([ET-VK][10/n] copy node, aten.repeat #3299)
add buck2 installation into setup.md
register view, reshape and select
Update llama2 readme file - main branch (Update llama2 readme file - main branch #3340)
Build custom ops in pybinding (Build custom ops in pybinding #3263)
Enable doc job to run on -rc tags. (Enable doc job to run on -rc tags. #3345)
Eliminate deprecated api usage (Eliminate deprecated api usage #2695)
Remove unneeded _to_copy in edge dialect.
Extend setup cmake ability (Extend setup cmake ability #3349)
Half support for index op
Add EXECUTORCH_SEPARATE_FLATCC_HOST_PROJECT cmake option (Add EXECUTORCH_SEPARATE_FLATCC_HOST_PROJECT cmake option #3356)
Export the ET_VERSION_DOCS variable in doc build (Export the ET_VERSION_DOCS variable in doc build #3358)
Fix extension/data_loader installation (Fix extension/data_loader installation #3355)
Reword "preview release" notice now that we are at alpha (Reword "preview release" notice now that we are at alpha #3364)
Fix quantized_linear cpp op schema
Add Disclaimer

Summary: Pull Request resolved: #2944 Need change from D55354487 to get mutable buffer + pt2e working Reviewed By: JacobSzwejbka Differential Revision: D55922254 fbshipit-source-id: 5ea4471eb0e22149a0dbb4e921fe447cceb13bf1

Summary: Pull Request resolved: #2883 ## Summary (cases handled) We introduce support for the convolution cases covered by ATen-VK's transpose implementation. This is achieved by - reusing the existing [`conv_transpose2d.glsl`](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/glsl/conv_transpose2d.glsl), and - [moving special weights prepacking from CPU](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/ops/Convolution.cpp#L134-L235) to the GPU in `conv_transpose2d_prepack_weights.glsl`. We also include resizing support for dynamic shapes. Note that only height and width of the input can vary. ## Cases not handled The implementation is on-par with ATen-VK's Transpose. This means the following cases are missing: 1. **Groups G > 1.** 2. **Batch (input) N > 1.** 3. **Dilation > 1.** ghstack-source-id: 221721754 exported-using-ghexport bypass-github-export-checks Reviewed By: copyrightly, SS-JIA Differential Revision: D55667336 fbshipit-source-id: 3b7b7c912ef947610624e2e1c5b753de393234a0

Summary: Pull Request resolved: #2884 ## Summary We introduce support for the convolution cases covered by [ATen-VK's default Depthwise implementation](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/ops/Convolution.cpp#L68). This is achieved by - reusing the [existing `conv2d_dw.glsl`](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/glsl/conv2d_dw.glsl), and - [moving special weights prepacking from CPU](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/ops/Convolution.cpp#L80-L132) to the GPU in `conv2d_dw_prepack_weights.glsl`. The implementation is on-par with ATen-VK's Depthwise. This means it only covers: - `in_channels == groups`, `out_channels == groups` A full implementation would cover, for any positive integer K: - `in_channels == groups`, `out_channels == groups * K` ghstack-source-id: 221721752 exported-using-ghexport bypass-github-export-checks Reviewed By: SS-JIA Differential Revision: D55813511 fbshipit-source-id: c0726798bd36cc5ff2326836c28a5f7d23494f5e

Summary: Pull Request resolved: #2854 ## Context Currently, when executing a `ComputeGraph` with prepacked tensors with [Vulkan Validation Layers](https://github.com/KhronosGroup/Vulkan-ValidationLayers) turned on, the following Validation Errors can be observed. Note that Validation Layers can be turned on by running Vulkan binaries on Mac with the `vkconfig` app opened. ``` UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout(ERROR / SPEC): msgNum: 1303270965 - Validation Error: [ UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout ] Object 0: handle = 0x7fb76dbbf988, type = VK_OBJECT_TYPE_COMMAND_BUFFER; | MessageID = 0x4dae5635 | vkQueueSubmit(): pSubmits[0].pCommandBuffers[0] command buffer VkCommandBuffer 0x7fb76dbbf988[] expects VkImage 0xd79c8a0000000f09[] (subresource: aspectMask 0x1 array layer 0, mip level 0) to be in layout VK_IMAGE_LAYOUT_GENERAL--instead, current layout is VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL. Objects: 1 [0] 0x7fb76dbbf988, type: 6, name: NULL ``` The reason for this is that prepacked textures are written to with `WRITE` memory access during packing, which means they will be in the `VK_IMAGE_LAYOUT_GENERAL` layout. However, they will subsequently be read from during `graph.execute()`, meaning the texture will have transitioned to `VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL`, but will be bound using the `VK_IMAGE_LAYOUT_GENERAL` layout. Subsequent calls to `execute()` will therefore see that the prepacked texture has been bound with the wrong layout, since after the first graph execution the texture will have the `VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL` layout. The solution is to submit a no-op shader dispatch during prepacking to trigger a transition to the `READ_ONLY_OPTIMAL` layout. ghstack-source-id: 221871426 bypass-github-pytorch-ci-checks Reviewed By: jorgep31415 Differential Revision: D55772003 fbshipit-source-id: f9c69e6e571ca0d0d28a6c25716766af98e82d41

…ut`s (#2948) Summary: Pull Request resolved: #2948 ## Context Introduce the following convenience `constexpr`: * `api::kBuffer`, `api::kTexture3D`, and `api::kTexture2D` * `api::kWidthPacked`, `api::kHeightPacked`, and `api::kChannelsPacked` Also remove the `api::StorageType::UNKNOWN` enum entry as it doesn't really serve any purpose. ghstack-source-id: 221871428 bypass-github-pytorch-ci-checks Reviewed By: copyrightly, jorgep31415 Differential Revision: D55811278 fbshipit-source-id: 26dc1706ac2605c13f247d08a21863ff3ef94488

Summary: Pull Request resolved: #2949 It is supposed to be unlikely for assert/check conditions to fail; let's tell the compiler about that. Reviewed By: mergennachin Differential Revision: D55929730 fbshipit-source-id: 5677c19cd8342cbd77a9c0b973059ed3d5ee800b

Summary: Pull Request resolved: #2772 Just a spelling mistake. Reviewed By: JacobSzwejbka Differential Revision: D55542731 fbshipit-source-id: c12bcab53661561bf0d8223d5cae9ed92b39e599

Summary: Pull Request resolved: #2773 Noticed this page didn't line up right. Now it does. Reviewed By: mergennachin, kirklandsign Differential Revision: D55542836 fbshipit-source-id: a25a376ce9e77f3bc360e9ab6cf15c9ae9ecc7bf

Summary: Pull Request resolved: #2885 We port an optimization from ATen-VK for specific weight sizes: [`conv2d_dw_output_tile.glsl`](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/glsl/conv2d_dw_output_tile.glsl) ghstack-source-id: 221887576 exported-using-ghexport bypass-github-export-checks Reviewed By: SS-JIA Differential Revision: D55814588 fbshipit-source-id: 86a85d122abbcebfed41466bc0a4907a6ddc80f9

Summary: Pull Request resolved: #2886 We port an optimization from ATen-VK for specific weight sizes: [`conv2d_pw.glsl`](https://github.com/pytorch/pytorch/blob/09c72eaa3f69f90402c86a30abf4fc621298578c/aten/src/ATen/native/vulkan/glsl/conv2d_pw.glsl) ghstack-source-id: 221887670 exported-using-ghexport bypass-github-export-checks Reviewed By: SS-JIA Differential Revision: D55814587 fbshipit-source-id: 419d82ddcf2dce59b2d1ec5abf313356fce074e6

Summary: Minor updates to the prerequisite section of the LLM getting started guide. Passing -s to pyenv install prevents a prompt if python 3.10 is already installed (it will just silently continue in this case when the flag is passed). Additionally, under pyenv, we should be using python, not python3. I also added a little bit of wording on env management. Pull Request resolved: #2940 Test Plan: Ran LLM guide prerequisite section on an m1 mac with pyenv-virtualenv. Reviewed By: byjlw Differential Revision: D55913382 Pulled By: GregoryComer fbshipit-source-id: 7f04262b025db83b8621c972c90d3cdc3f029377

Summary: Version hash reported by https://github.com/facebook/buck2/releases/download/2024-02-15/buck2-x86_64-apple-darwin.zst Pull Request resolved: #2868 Reviewed By: Olivia-liu Differential Revision: D55914146 Pulled By: GregoryComer fbshipit-source-id: b9882900acfd4cb6f74eda90a7c99bdb119ec122

Summary: Pull Request resolved: #2825 Add capability to print the node list with arguments to allow better debugging. Reviewed By: SS-JIA Differential Revision: D55510335 fbshipit-source-id: 151e3a6f249417dfe644172c1b5f0e83a3b110dd

Summary: Pull Request resolved: #2887 The final touches to get ET-VK convolution on-par with ATen-VK's convolution. ## Idea In our shaders, we add the bias to our sum. ``` ${VEC4_T[DTYPE]} sum = texelFetch(bias_in, ivec2(pos.z, 0), 0); ``` To keep our shaders as is, we implement having no bias by allocating a buffer of zeros. Then, our shader adds zero to our sum. ## Issue If `Bias=False`, dummy buffer of zeros is not serialized with the graph. The bias ValueRef is deserialized in the runtime as `TypeTag::NONE`, not `TypeTag::TENSORREF`. ## Solution If `TypeTag::NONE` is given, (1) create the `vTensor` using the `out_channels` value from the weights, (2) allocate a StagingBuffer of that size, and (3) `memset` its data to zero. Failure to do (3) will result in undefined behavior. ghstack-source-id: 221926167 exported-using-ghexport bypass-github-export-checks Reviewed By: SS-JIA Differential Revision: D55814589 fbshipit-source-id: ce7b82c31bb11540ed2d98ab14131841fcee93e4

Summary: Pull Request resolved: #2920 TSIA ghstack-source-id: 221926168 exported-using-ghexport bypass-github-export-checks Reviewed By: SS-JIA Differential Revision: D55829466 fbshipit-source-id: 48b4f15c41141093dd061c43e6b769eb4c25c81b

Summary: Pull Request resolved: #2807 The operator `aten.sum.dim_IntList` could take an empty list as the parameter for `dims`. We modify `vulkan_graph_builder.py` to accommodate the empty list. Moreover, the op `aten.sum.default` is implemented as a [decomposition](https://www.internalfb.com/code/fbsource/[96e496f9db8f92967b4394bd4f60e39ab916740b]/xplat/caffe2/torch/_decomp/decompositions.py?lines=4676) into `aten.sum.dim_IntList` with empty `dims`. So we will support `aten.sum.default` with the changes. Context: `torch.sum(x, ())` and `torch.sum(x)` are two ways to compute the sum of all elements in tensor `x`. Reviewed By: SS-JIA, jorgep31415 Differential Revision: D55630993 fbshipit-source-id: 923d276118e893ff6885b92eb7b4c7cb7a95b374

Summary: Pull Request resolved: #2961 Fix these 3 CI job failures caused by #2934 (D55907752): * Apple / build-frameworks-ios / macos-job * trunk / test-arm-backend-delegation / linux-job * trunk / test-coreml-delegate / macos-job Reviewed By: kirklandsign Differential Revision: D55950023 fbshipit-source-id: 6166d9112e6d971d042df1400442395d8044c3b3

Summary: Pull Request resolved: #2964 ## Context Some research into efficient string concatenation suggests that streams in C++ are not quite efficient. The best way to concatenate strings seems to be creating a `std::string` and reserving sufficient capacity for the `std::string`. This diff deprecates the usage of `std::stringstream` when constructing kernel names in favor of using `std::string` directly. Reviewed By: copyrightly Differential Revision: D55951475 fbshipit-source-id: a1a584669e80984b85d11b7d6d4f7593290e562b

) Summary: Pull Request resolved: #2952 * Some auto-formatting by my VSCode (remove extra spaces) * Remove imports that have been imported in previous part of the doc * Other minor changes to keep consistency across the doc * Link a screenshot instead of using the raw table because the original table is illegible: {F1482781056} Reviewed By: GregoryComer Differential Revision: D55938344 fbshipit-source-id: 699abb9ebe1196ab73d90a3d08d60be7aa0d8688

Summary: * Update tutorial due to recent changes. * Clean up setup.sh for app helper lib build. Pull Request resolved: #2962 Reviewed By: cccclai Differential Revision: D55951189 Pulled By: kirklandsign fbshipit-source-id: 2c95e8580145b039f503e7cd99a4003867f8dbb0

Summary: - Add per channel weight quantization for linear op - Bias quantization for per channel weight Linear op is not support yet Pull Request resolved: #2822 Reviewed By: kirklandsign Differential Revision: D55731629 Pulled By: cccclai fbshipit-source-id: 831a47c897b34e1a749325df56a8bbd0acda80e1

Summary: Pull Request resolved: #2936 Fix the way we use `at::from_blob()` and add proper namespace to `CompileTimeFunctionPointer` so to not confused with `at::CompileTimeFunctionPointer`. bypass-github-pytorch-ci-checks bypass-export-ci-checks Reviewed By: lucylq Differential Revision: D55907751 fbshipit-source-id: ad793e30ec72f48e7300d75820209035d42cae6c

Summary: Pull Request resolved: #2935 Currently `EXECUTORCH_BUILD_CUSTOM` is not being respected properly. If this option is false, we should not build `llama2/custom_ops` anywhere. If this option is true, we should build `llama2/custom_ops` in both llama runner binary and pybind. This PR consolidates it. bypass-github-pytorch-ci-checks bypass-export-ci-checks Reviewed By: lucylq Differential Revision: D55907750 fbshipit-source-id: 03a7a8cbd499c734060de385d6edb193cf35470d

Summary: Pull Request resolved: #2954 Change the tokenizer APIs to: ``` Result<std::vector<uint64_t>> encode(const std::string& input, int8_t bos, int8_t eos); Result<std::string> decode(uint64_t prev_token, uint64_t token); ``` Notice that: we use `uint64_t` for token id just to be safe. We return a std::vector of tokens for encode() API. Reviewed By: lucylq Differential Revision: D55944780 fbshipit-source-id: 9b44437e7061424526f4e0b049a3449129f0ba53

Summary: Pull Request resolved: #2033 Update the OSS Xtensa repo with more up to date compiler and quantizer things. Introduce a test folder and a conv1d test. Reviewed By: tarun292, cccclai Differential Revision: D54034581 fbshipit-source-id: c2bf0c43897a2ef7dff291698370d2583433a6ba

…LLM (#2977) Summary: Pull Request resolved: #2977 As titled Reviewed By: Gasoonjia Differential Revision: D55992093 fbshipit-source-id: 7864c330bd86af5d4127cacfd47e96f1e6666bfb

Summary: Pull Request resolved: #2981 As titled, a quick follow up of D55907750 Reviewed By: lucylq Differential Revision: D55996735 fbshipit-source-id: f535b013b7b900c5a2c2ed79f6b6738dcf1f91ec

Summary: After pytorch/test-infra#5086, the working directory is now set correctly, so `pushd` isn't needed anymore. More importantly, trying to change the directory ends up failing all macOS CI jobs because that subdirectory doesn't exist. Pull Request resolved: #2980 Reviewed By: larryliu0820 Differential Revision: D55996299 Pulled By: huydhn fbshipit-source-id: 05758603d7628cc0a01fd577a49202d45c84e6c5

Summary: I'm trying to setup a simple perf test when running llama2 on Android. It's naively sent a prompt and record the TPS. Open for comment about the test here before setting this up on CI. ### Testing Copy the exported model and the tokenizer as usual, then cd to the app and run `./gradlew :app:connectAndroidTest`. The test will fail if the model is failed to load or if the TPS is lower than 7 as measure by https://github.com/pytorch/executorch/tree/main/examples/models/llama2 Pull Request resolved: #2963 Reviewed By: kirklandsign Differential Revision: D55951637 Pulled By: huydhn fbshipit-source-id: 34c189aefd7e31514fcf49103352ef3cf8e5b2c9

Summary: It was a workaround to skip `aten.index_put` op in Core ML delegation, at the cost of partitioning the Llama model into 13 pieces. For better performance, we prefer to delegate the whole model to Core ML. Since Core ML has added the [necessary support](apple/coremltools#2190), it is time to revert this workaround Pull Request resolved: #2975 Reviewed By: kirklandsign Differential Revision: D56002979 Pulled By: cccclai fbshipit-source-id: e7a7c8c43706cb57eba3e6f720b3d713bec5065b

Summary: Use relative markdown links instead of full URLs. This way, the docs will always point to a consistent branch. Pull Request resolved: #3244 Test Plan: Clicked on all modified links in the rendered docs preview: https://docs-preview.pytorch.org/pytorch/executorch/3244/llm/getting-started.html Reviewed By: Gasoonjia Differential Revision: D56479234 Pulled By: dbort fbshipit-source-id: 45fb25f017c73df8606c3fb861acafbdd82fec8c

Summary: - Added Llama 3 8B - Added llm_manual in the list - changed name from Extensa to Cadence Pull Request resolved: #3275 Reviewed By: Gasoonjia Differential Revision: D56524960 Pulled By: iseeyuan fbshipit-source-id: 2b4464028fe3cdf3c2b524d233fa3e87b2561dda

Differential Revision: D56480274 Original commit changeset: 451116a0f907 Original Phabricator Diff: D56480274 fbshipit-source-id: e9603e5076113560b1224a56432abf321f82e284

Summary: Pull Request resolved: #3300 This diff solves part of Ali's comments in our tracer sheet (https://docs.google.com/spreadsheets/d/1PoJt7P9qMkFSaMmS9f9j8dVcTFhOmNHotQYpwBySydI/edit#gid=0). Specifically speaking： "NanoGPT" -> "nanoGPT" "CoreML" -> "Core ML" "ExecuTorch Codebase" -> "ExecuTorch codebase" "Android Phone" -> "Android phone" "How to build Mobile Apps" -> "How to Build Mobile Apps" also shorten the following two column names for avoid overlapping. "occurrences_in_delegated_graphs" -> "# in_delegated_graphs" "occurrences_in_non_delegated_graphs" -> # in_non_delegated_graphs Reviewed By: Jack-Khuu Differential Revision: D56513601 fbshipit-source-id: 7015c2c5b94b79bc6c57c533ee812c9e58ab9d56

Summary: . Reviewed By: cccclai Differential Revision: D56532283 fbshipit-source-id: 62d7c9e8583fdb5c9a1b2e781e80799c06682aae

Summary: As titled Reviewed By: lucylq, Gasoonjia, guangy10 Differential Revision: D56532035 fbshipit-source-id: ddf4f3864db0f200b97e67673a7086dac790eb82

Summary: - add note for embedding quantize, for llama3 - re-order export args to be the same as llama2, group_size missing `--` Pull Request resolved: #3315 Reviewed By: cccclai Differential Revision: D56528535 Pulled By: lucylq fbshipit-source-id: 4453070339ebdb3d782b45f96fe43d28c7006092

Summary: Pull Request resolved: #3298 Reviewed By: Olivia-liu Differential Revision: D56509749 Pulled By: tarun292 fbshipit-source-id: 36b56e7cc039144105d64431697a16a793029af8

Summary: . Reviewed By: cccclai Differential Revision: D56535633 fbshipit-source-id: 070a3b0af9dea234f8ae4be01c37c03b4e0a56e6

…ner (#3324) Summary: **Summary of changes**: - Update MPS documentation to reflect all changes since previous release - Add helper script to build `mps_executor_runner` **Testing**: - Verified that mps_executor_runner builds correctly: ``` ./examples/apple/mps/scripts/build_mps_executor_runner.sh /examples/apple/mps/scripts/build_mps_executor_runner.sh --Debug ``` Verified that the docs are building correctly: ``` cd docs make html ``` cc shoumikhin, cccclai Pull Request resolved: #3324 Reviewed By: shoumikhin Differential Revision: D56535774 Pulled By: cccclai fbshipit-source-id: 5974795732dbe1089e3d63cd1b618cadf7a2573e

… as Custom Metal kernels are yet not enabled) (#3328) Summary: Remove the sorting of the nodes from partitioning (not needed for now as Custom Metal kernels are yet not enabled) **Testing**: Verified that tracing works correctly with release branch: `python3 -m examples.apple.mps.scripts.mps_example --model_name="mv3"` cc shoumikhin , cccclai Pull Request resolved: #3328 Reviewed By: shoumikhin Differential Revision: D56540389 Pulled By: cccclai fbshipit-source-id: e8a53f624b58ac4d2348c87e08acd5f2fb3de5b2

Summary: Pull Request resolved: #3299 1. Introduce a `CopyNode` for generic copy-with-offset operations. 2. `aten.repeat` on all dimensions. 2.1 Use `CopyNode` where possible. 2.2. Specialized `repeat_channel` shader to handle packings 3. Update codegen to support `Methods` variant only operations. Need a new route to trigger the dispatch. ghstack-source-id: 223812048 Reviewed By: copyrightly Differential Revision: D56499329 fbshipit-source-id: 72936e621940588ce398dd62669ec9aa637e98ba

Summary: bring buck2 installation back, and scrub any "-DBUCK2=buck2" in our docs, to unblock users from using buck2 Reviewed By: guangy10 Differential Revision: D56540769 fbshipit-source-id: 363e592c17dd2747a693e59d8d6b6d20f43c8451

Summary: - We register `select`, `unsqueeze` and `view` in `vulkan_partitioner.py` in order to run vulkan_delegate test (Python e2e test). The latter two might be used to implement `bmm` and `addmm`, so I want to make sure they work. - We register `reshape` in `View.cpp` explicitly. `reshape` is implemented through `_reshape_alias` (see [this](https://www.internalfb.com/code/fbsource/[a3dd6401f00d73f09bbdea63887fef54ea2c6dd2]/fbcode/caffe2/aten/src/ATen/native/native_functions.yaml?lines=4872-4881)) which is [decomposed as `view`](https://www.internalfb.com/code/fbsource/[bbb783ae1cff98b3b549da3edd845dde946d3da8]/xplat/caffe2/torch/_decomp/decompositions.py?lines=3669-3672). For codegen test, we still need to register the op, otherwise there is error ``` C++ exception with description "Exception raised from get_op_fn at xplat/executorch/backends/vulkan/runtime/graph/ops/OperatorRegistry.cpp:20: (it != table_.end()) is false! Could not find operator with name aten.reshape.default" thrown in the test body. ``` Reviewed By: yipjustin, liuk22 Differential Revision: D56454941 fbshipit-source-id: c83e6fb97d9cf9019cc6e786508f353a22236931

Summary: Pull Request resolved: #3340 Reviewed By: orionr, kimishpatel, cccclai Differential Revision: D56553088 Pulled By: mergennachin fbshipit-source-id: 2994dd3ab2692c5b972316af1617bd06d647af96

Summary: Right now we are not building it and it is causing missing ops in torchchat. This PR adds it into pybinding. Pull Request resolved: #3263 Reviewed By: lucylq Differential Revision: D56500693 Pulled By: larryliu0820 fbshipit-source-id: 0ed0e28fcccb6002ef48e6a38b60e92d8af4def6

Summary: Pull Request resolved: #3345 Reviewed By: dbort Differential Revision: D56557091 Pulled By: svekars fbshipit-source-id: 4300ca86d01ec110fc6934588cd691c12661a730

Summary: Pull Request resolved: #2695 Reviewed By: mergennachin Differential Revision: D55091814 fbshipit-source-id: 04b2a888c6bbdaa195cb916c6564aa93daca2514

Summary: In executorch we will dtype-specialize the kernels and also run on a single device with export. Therefore _to_copy is not needed in edge dialect. Reviewed By: tugsbayasgalan Differential Revision: D56579169 fbshipit-source-id: 5a2e3cd453a11bd2ad009b439587b0fc589f7fe4

Summary: For executorch users, we see a common pattern that they have to: ```bash bash install_requirements.sh --pybind xnnpack cmake -S . -Bcmake-out ... cmake --build ... ``` This is repeating cmake build twice, the first one is inside setup.py. Here I'm adding a way to allow setup.py to install the libraries seperately, by passing `CMAKE_ARGS` and `CMAKE_BUILD_ARGS` into setup.py, through `install_requirements.sh`. After this change, user can do: ```bash export CMAKE_ARGS="-DCMAKE_INSTALL_PREFIX=<install dir> \ -DEXECUTORCH_BUILD_OPTIMIZED=ON \ ..." export CMAKE_BUILD_ARGS="--target install" bash install_requirements.sh --pybind xnnpack ``` Then we should be able to find `libxnnpack.a` `liboptimized_ops_lib.a` etc under install dir. Pull Request resolved: #3349 Reviewed By: mikekgfb Differential Revision: D56560786 Pulled By: larryliu0820 fbshipit-source-id: fb6cd230df2317067f07ae0f1e72d0596b7b454b

Reviewed By: cccclai Differential Revision: D56543186 fbshipit-source-id: 4fed6b9b3ede3cdcb67a9a52150e3f22cc02b180

Summary: Currently, we always build two copies of the flatcc targets, just in case we happen to be cross-compiling. But because the flatcc project puts its binaries in the source directory, those two copies can interfere with each other. We don't need to build two copies when not cross-compiling, so add a new option to avoid the second "host" build. Eventually we should only enable this when cross-compiling, but for now disable it when building the pip package (which is never cross-compiled). Pull Request resolved: #3356 Test Plan: `rm -rf pip-out && ./install_requirements.sh` succeeded. Looking in the `pip-out/temp.*/cmake-out` directory, there is no `_host_build` directory, but the etdump headers were successfully generated under `pip-out/temp.*/cmake-out/sdk/include/executorch/sdk/etdump/`. Reviewed By: malfet, larryliu0820 Differential Revision: D56582507 Pulled By: dbort fbshipit-source-id: 4ce6c680657bc57cfcf016826364a3f46c4c953e

Summary: Pull Request resolved: #3358 Reviewed By: dbort Differential Revision: D56584847 Pulled By: svekars fbshipit-source-id: 77c4105edf15503bf1b29c1f120111a73b973c4c

Summary: `libextension_data_loader.a` is not installed properly. This PR removes the prefix so that it can be properly installed Pull Request resolved: #3355 Test Plan: See `libextension_data_loader.a` showing up under executorch/cmake-out/lib. Reviewed By: lucylq, mikekgfb Differential Revision: D56580943 Pulled By: larryliu0820 fbshipit-source-id: b771192d03799fd576e8591ec7c45fae23f20762

Summary: Pull Request resolved: #3364 Test Plan: https://docs-preview.pytorch.org/pytorch/executorch/3364/index.html Reviewed By: svekars Differential Revision: D56596949 Pulled By: dbort fbshipit-source-id: f6c71e072bcefbb7d04354d1ef78d780c14facb5

Summary: The cpp op schema does not match the registered one. Fix that. Reviewed By: tarun292, cccclai Differential Revision: D56594373 fbshipit-source-id: cb4853030715245e7a0177c0f193c4558f19584d

pytorch-bot · 2024-04-26T18:28:15Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/3376

📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

cccclai and others added 30 commits April 9, 2024 13:28

update torch pin (#2944)

cb6ddae

Summary: Pull Request resolved: #2944 Need change from D55354487 to get mutable buffer + pt2e working Reviewed By: JacobSzwejbka Differential Revision: D55922254 fbshipit-source-id: 5ea4471eb0e22149a0dbb4e921fe447cceb13bf1

s/heirarchies/hierarchies/ (#2772)

3661a11

Summary: Pull Request resolved: #2772 Just a spelling mistake. Reviewed By: JacobSzwejbka Differential Revision: D55542731 fbshipit-source-id: c12bcab53661561bf0d8223d5cae9ed92b39e599

Fix indentation in selective build example code (#2773)

02f565e

Summary: Pull Request resolved: #2773 Noticed this page didn't line up right. Now it does. Reviewed By: mergennachin, kirklandsign Differential Revision: D55542836 fbshipit-source-id: a25a376ce9e77f3bc360e9ab6cf15c9ae9ecc7bf

Compute graph print readable (#2825)

564c276

Summary: Pull Request resolved: #2825 Add capability to print the node list with arguments to allow better debugging. Reviewed By: SS-JIA Differential Revision: D55510335 fbshipit-source-id: 151e3a6f249417dfe644172c1b5f0e83a3b110dd

Add convolution cases to codegen (#2920)

f0bfc3c

Summary: Pull Request resolved: #2920 TSIA ghstack-source-id: 221926168 exported-using-ghexport bypass-github-export-checks Reviewed By: SS-JIA Differential Revision: D55829466 fbshipit-source-id: 48b4f15c41141093dd061c43e6b769eb4c25c81b

Add the missing import generate_etrecord to doc Getting Started with …

cb9caa3

…LLM (#2977) Summary: Pull Request resolved: #2977 As titled Reviewed By: Gasoonjia Differential Revision: D55992093 fbshipit-source-id: 7864c330bd86af5d4127cacfd47e96f1e6666bfb

Fix llama runner test (#2981)

75c27c3

Summary: Pull Request resolved: #2981 As titled, a quick follow up of D55907750 Reviewed By: lucylq Differential Revision: D55996735 fbshipit-source-id: f535b013b7b900c5a2c2ed79f6b6738dcf1f91ec

dbort and others added 27 commits April 24, 2024 09:32

Revert D56480274: Add more prebuilt artifacts

727a68d

Differential Revision: D56480274 Original commit changeset: 451116a0f907 Original Phabricator Diff: D56480274 fbshipit-source-id: e9603e5076113560b1224a56432abf321f82e284

Update readme.

ce1e9c1

Summary: . Reviewed By: cccclai Differential Revision: D56532283 fbshipit-source-id: 62d7c9e8583fdb5c9a1b2e781e80799c06682aae

Update custom kernel registration API

f6758fc

Summary: As titled Reviewed By: lucylq, Gasoonjia, guangy10 Differential Revision: D56532035 fbshipit-source-id: ddf4f3864db0f200b97e67673a7086dac790eb82

Fix sdk_example_runner.sh (#3298)

aa3e736

Summary: Pull Request resolved: #3298 Reviewed By: Olivia-liu Differential Revision: D56509749 Pulled By: tarun292 fbshipit-source-id: 36b56e7cc039144105d64431697a16a793029af8

Update readme.

035aee4

Summary: . Reviewed By: cccclai Differential Revision: D56535633 fbshipit-source-id: 070a3b0af9dea234f8ae4be01c37c03b4e0a56e6

add buck2 installation into setup.md

590cbce

Summary: bring buck2 installation back, and scrub any "-DBUCK2=buck2" in our docs, to unblock users from using buck2 Reviewed By: guangy10 Differential Revision: D56540769 fbshipit-source-id: 363e592c17dd2747a693e59d8d6b6d20f43c8451

Update llama2 readme file - main branch (#3340)

79b79cb

Summary: Pull Request resolved: #3340 Reviewed By: orionr, kimishpatel, cccclai Differential Revision: D56553088 Pulled By: mergennachin fbshipit-source-id: 2994dd3ab2692c5b972316af1617bd06d647af96

Enable doc job to run on -rc tags. (#3345)

fd63d0c

Summary: Pull Request resolved: #3345 Reviewed By: dbort Differential Revision: D56557091 Pulled By: svekars fbshipit-source-id: 4300ca86d01ec110fc6934588cd691c12661a730

Eliminate deprecated api usage (#2695)

8fcba36

Summary: Pull Request resolved: #2695 Reviewed By: mergennachin Differential Revision: D55091814 fbshipit-source-id: 04b2a888c6bbdaa195cb916c6564aa93daca2514

Half support for index op

7b3b485

Reviewed By: cccclai Differential Revision: D56543186 fbshipit-source-id: 4fed6b9b3ede3cdcb67a9a52150e3f22cc02b180

Export the ET_VERSION_DOCS variable in doc build (#3358)

c32b0a2

Summary: Pull Request resolved: #3358 Reviewed By: dbort Differential Revision: D56584847 Pulled By: svekars fbshipit-source-id: 77c4105edf15503bf1b29c1f120111a73b973c4c

Fix quantized_linear cpp op schema

44d4bac

Summary: The cpp op schema does not match the registered one. Fix that. Reviewed By: tarun292, cccclai Differential Revision: D56594373 fbshipit-source-id: cb4853030715245e7a0177c0f193c4558f19584d

Add Disclaimer

3fe25df

facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Apr 26, 2024

mergennachin closed this Apr 26, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

disclaimer #3376

disclaimer #3376

mergennachin commented Apr 26, 2024

pytorch-bot bot commented Apr 26, 2024

disclaimer #3376

disclaimer #3376

Conversation

mergennachin commented Apr 26, 2024

pytorch-bot bot commented Apr 26, 2024

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/3376