Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Weekly issue metrics report - 2025-02-01..2025-02-07 #7

Open
github-actions bot opened this issue Feb 17, 2025 · 0 comments
Open

Weekly issue metrics report - 2025-02-01..2025-02-07 #7

github-actions bot opened this issue Feb 17, 2025 · 0 comments
Assignees
@wdvr

Comments

@github-actions
Copy link

Issue Metrics

Metric Average Median 90th percentile
Time to first response 1 day, 21:07:37 6:26:09 5 days, 19:26:22
Time to close 3 days, 20:23:20 5 days, 9:51:01 7 days, 5:59:36
Time to answer None None None
Time spent in module: android 11 days, 2:50:05 10 days, 22:34:32 12 days, 3:48:53
Time spent in module: ios 7 days, 20:28:23 7 days, 5:26:23 10 days, 4:12:09
Time spent in module: benchmark 11 days, 8:10:51 11 days, 7:44:51 11 days, 8:04:23
Time spent in module: build/install 10 days, 15:39:54 10 days, 20:17:57 13 days, 0:00:55
Time spent in module: vulkan 5 days, 19:28:27 5 days, 19:28:27 9 days, 11:53:22
Time spent in module: llm 13 days, 1:16:12 13 days, 1:16:12 14 days, 7:10:06
Time spent in module: ci 9 days, 10:14:05 9 days, 10:14:05 12 days, 15:26:10
Time spent in module: training 9 days, 16:49:38 9 days, 16:49:38 11 days, 16:20:00
Time spent in module: user experience None None None
Metric Count
Number of items that remain open 45
Number of items closed 11
Number of most active mentors 0
Total number of items created 56
Title URL Author Time to first response Time to close Time to answer Time spent in module: android Time spent in module: ios Time spent in module: benchmark Time spent in module: build/install Time spent in module: vulkan Time spent in module: llm Time spent in module: ci Time spent in module: training Time spent in module: user experience
Add executorch & pytorch to godbolt.org pytorch/executorch#8324 awgr None 2 days, 16:53:01 None None None None None None None None None None
Installing from source on Google Cloud Ubuntu 24.04 pytorch/executorch#8311 SFrav 0:33:14 5 days, 19:42:10 None None None None 5 days, 19:08:44 None None None None None
Feature Request: Integrate MarkdownView in Android Demo App for Proper Markdown Rendering pytorch/executorch#8309 JamePeng None None None 9 days, 15:12:58.725501 None None None None None None None None
Benchmark Hugging Face models via Optimum integration pytorch/executorch#8294 guangy10 None None None None None 10 days, 9:49:25.915047 None None None None None None
Add Vulkan Delegate Benchmarks for Llama 3.2 (1B and 3B) and MobileNet pytorch/executorch#8293 SS-JIA None None None None None 10 days, 9:59:39.148858 None 10 days, 9:59:36.148865 None None None None
Support Python 3.13 pytorch/executorch#8292 GregoryComer 21:21:42 None None None None None 10 days, 10:12:58.425501 None None None None None
Add GenerateFromPoS in Android LLAMA API pytorch/executorch#8290 kirklandsign 1:12:06 None None 10 days, 11:02:31.708346 None None None None None None None None
Provide an example model in the getting started guide pytorch/executorch#8283 GregoryComer 1:51:07 None None None None None None None None None None None
Advise on how to run the training example on iOS pytorch/executorch#8282 YuanTingHsieh 0:54:22 7 days, 5:59:36 None None 7 days, 5:26:23 None None None None None 7 days, 5:26:39 None
Move examples + tests to single-step CMake build pytorch/executorch#8274 GregoryComer None None None None None None 10 days, 19:43:58.599916 None None None None None
Improve common error messages pytorch/executorch#8273 GregoryComer 1 day, 8:16:57 None None None None None None None None None None None
Support runtime configuration pytorch/executorch#8272 GregoryComer None 4:07:09 None None None None None None None None None None
Pull iOS docs up into the top-level getting started flow pytorch/executorch#8271 GregoryComer 4:07:14 None None None None None None None None None None None
Re-write export tutorial / export getting started pytorch/executorch#8270 GregoryComer 4:11:51 None None None None None None None None None None None
Create delegate documentation template pytorch/executorch#8269 GregoryComer 4:20:02 None None None None None None None None None None None
Remove buck dependency from CMake build pytorch/executorch#8268 GregoryComer 1 day, 12:38:48 None None None None None 10 days, 20:51:55.783608 None None None None None
Provide API to query backend availability at runtime pytorch/executorch#8267 GregoryComer 4:32:34 None None None None None None None None None None None
Enable XNNPACK workspace sharing by default pytorch/executorch#8266 GregoryComer 5:35:47 6:27:24 None None None None None None None None None None
Support per-op partitioning in XNNPACK delegate for NHWC ops pytorch/executorch#8265 GregoryComer 7 days, 8:32:41 7 days, 8:32:52 None None None None None None None None None None
Document iOS Version Requirement for Core ML Delegate pytorch/executorch#8264 GregoryComer 1 day, 12:14:17 None None None 10 days, 21:53:36.094415 None None None None None None None
Swift and/or Objective-C Bindings pytorch/executorch#8263 GregoryComer 5:31:33 5 days, 10:05:11 None None 5 days, 10:05:11 None None None None None None None
API reference for Android Java bindings pytorch/executorch#8262 GregoryComer None None None 10 days, 22:20:42.982488 None None None None None None None None
Standalone AAR build script pytorch/executorch#8261 GregoryComer None None None 10 days, 22:34:32.225899 None None None None None None None None
Expose runtime error codes / results in JNI pytorch/executorch#8260 GregoryComer None None None 10 days, 22:37:19.438635 None None None None None None None None
Distribute Android binaries through a Maven repository pytorch/executorch#8259 GregoryComer 7:13:52 None None 10 days, 22:59:23.708739 None None None None None None None None
Prototype benchmarking with private/custom devices that are not leased via AWS pytorch/executorch#8251 guangy10 5 days, 16:48:02 None None None None 11 days, 7:06:57.981599 None None None None None None
Support benchmark with arbitrary binaries on Android pytorch/executorch#8250 guangy10 None None None None None 11 days, 7:15:21.245400 None None None None None None
Support benchmark with arbitrary binaries on iOS pytorch/executorch#8249 guangy10 None None None None None 11 days, 7:16:18.540073 None None None None None None
DevX: Track the benchmark infra health and usage pytorch/executorch#8247 guangy10 5 days, 20:15:09 None None None None 11 days, 7:24:36.938024 None None None None None None
Support benchmark using prebuilt artifacts pytorch/executorch#8246 guangy10 5 days, 20:38:15 None None None None 11 days, 7:32:48.396580 None None None None None None
Support scheduling benchmark jobs via scripts pytorch/executorch#8245 guangy10 5 days, 1:40:21 None None None None 11 days, 7:44:50.740534 None None None None None None
Enable workflow time profiling pytorch/executorch#8244 guangy10 None None None None None 11 days, 7:48:40.023287 None None None None None None
Enable framegraphs for benchmarking on iOS pytorch/executorch#8243 guangy10 None None None None None 11 days, 7:53:49.313277 None None None None None None
Enable framegraphs for benchmarking on Android pytorch/executorch#8242 guangy10 5 days, 18:37:36 None None None None 11 days, 7:58:08.642060 None None None None None None
Integrate mem profiler to benchmark infra pytorch/executorch#8241 guangy10 None None None None None 11 days, 8:00:12.890120 None None None None None None
Integrate ETdump to benchmark infra pytorch/executorch#8240 guangy10 3:00:53 None None None None 11 days, 8:01:10.374207 None None None None None None
Support auto regression detection and bisect pytorch/executorch#8239 guangy10 5 days, 20:16:48 None None None None 11 days, 8:06:30.698463 None None None None None None
Kernels assume dim <= kTensorDimensionLimit pytorch/executorch#8237 dbort 18:22:23 None None None None None None None None None None None
Optimized linear op doesn't support bias pytorch/executorch#8234 swolchok 5 days, 14:01:04 None None None None None None None None None None None
[etLLM] Explore the way to load a new LLM to llama_transformer pytorch/executorch#8231 iseeyuan None None None None None None None None 11 days, 11:53:49.563806 None None None
ExecuTorch Vulkan Backend on Linux - Request for Documentation/Guidance pytorch/executorch#8211 csn1800 4:05:44 1 day, 4:57:20 None None None None None 1 day, 4:57:18 None None None None
Simple ConvNet causes mismatched dtypes during to_edge() call pytorch/executorch#8206 holgerroth 3:29:21 None None None None None None None None None 12 days, 4:12:36.148393 None
Linking errors when trying to use pre built executorch binaries for running lowered resnet50 in executorch runtime. pytorch/executorch#8196 sidharthkumar10500 3:07:22 None None None None None 12 days, 8:20:03.483357 None None None None None
CI timeout (test-llava-runner-linux) since pytorch#7922 pytorch/executorch#8180 swolchok 7:48:57 5 days, 9:51:01 None None None None None None None 5 days, 9:43:58 None None
Generating ETDump fails when using XNNPACK delegation pytorch/executorch#8177 spalatinate 4:06:16 None None None None None None None None None None None
Improve debuggability for kernel registration issues pytorch/executorch#8171 larryliu0820 None 6 days, 17:00:35 None None None None None None None None None None
Inconsistent logo between GitHub repo and documentation site pytorch/executorch#8163 guangy10 3 days, 23:17:16 None None None None None None None None None None None
High-Level Export API pytorch/executorch#8158 GregoryComer None 0:40:21 None None None None None None None None None None
Warn or error when using unsupported Python version pytorch/executorch#8156 GregoryComer None None None None None None None None None None None None
QLora and SpintQuant recipes fail to export on CI pytorch/executorch#8154 guangy10 5:38:26 None None None None 13 days, 10:44:11.761077 None None None 13 days, 10:44:11.761086 None None
Compile with -Wsign-compare flag pytorch/executorch#8149 mergennachin None None None None None None None None None None None None
QNN Partitioner - error with MatMul op pytorch/executorch#8139 metinsuloglu 5:00:37 None None None None None None None None None None None
MatMul quantisation fails with XNNPack backend pytorch/executorch#8138 metinsuloglu 3:07:53 None None None None None None None None None None None
Support ETDump Generation in JNI Bindings pytorch/executorch#8135 GregoryComer 2 days, 14:00:26 None None 13 days, 23:03:07.346820 None None None None None None None None
Support DeepSeek-VL2 models with MoE and MLA in ExecuTorch pytorch/executorch#8132 iseeyuan None None None None None None None None 14 days, 14:38:34.571145 None None None
Problem when try to install Executorh in Kaggle pytorch/executorch#8129 Ngapa 4 days, 18:03:27 None None None None None 13 days, 15:41:45.831323 None None None None None

This report was generated with the Issue Metrics Action
Search query used to find these items: repo:pytorch/executorch is:issue created:2025-02-01..2025-02-07
wdvr pushed a commit that referenced this issue Feb 28, 2025
@drisspg @pytorchmergebot
[Submodule] Turning flash-attention integration into 3rd party submod (…
3ecfe6b 
…pytorch#144120) (pytorch#146372)

Summary:

# Summary

### Sticky points

Cuda-graph rng handling has changed / deviated from original implementation. We will be left with a dangling 'offset' val and confusing naming due to BC

## Dependencies
- Flash PR: Dao-AILab/flash-attention#1419

### Other Points
- The BC linter is complaining about losing generate.py and its functions which is not real BC surface
cc albanD

imported-using-ghimport

Test Plan:
Imported from OSS

Building in dev
`buck build @//mode/dev-nosan -c fbcode.nvcc_arch=h100a  //caffe2:ATen-cu --show-full-output    `

I and Nming the .so I do see that the flash symbols are correctly named:
```
0000000001c3dfb0 t pytorch_flash::run_mha_bwd(pytorch_flash::Flash_bwd_params&, CUstream_st*)::$_0::operator()() const::{lambda()#1}::operator()() const::{lambda()#1}::operator()() const::{lambda()#7}::operator()() const
0000000001c36080 t pytorch_flash::run_mha_fwd(pytorch_flash::Flash_fwd_params&, CUstream_st*, bool)::$_0::operator()() const::{lambda()#2}::operator()() const::{lambda()#1}::operator()() const::{lambda()#6}::operator()() const
0000000001c360e0 t pytorch_flash::run_mha_fwd(pytorch_flash::Flash_fwd_params&, CUstream_st*, bool)::$_0::operator()() const::{lambda()#2}::operator()() const::{lambda()#1}::operator()() const::{lambda()#7}::operator()() const
0000000001c35fc0 t pytorch_flash::run_mha_fwd(pytorch_flash::Flash_fwd_params&, CUstream_st*, bool)::$_0::operator()() const::{lambda()#1}::operator()() const::{lambda()#1}::operator()() const::{lambda()#6}::operator()() const
0000000001c36020 t pytorch_flash::run_mha_fwd(pytorch_flash::Flash_fwd_params&, CUstream_st*, bool)::$_0::operator()() const::{lambda()#1}::operator()() const::{lambda()#1}::operator()() const::{lambda()#7}::operator()() const
```

Reviewed By: vkuzo

Differential Revision: D68502879

Pulled By: drisspg

Pull Request resolved: pytorch#146372
Approved by: https://github.com/jbschlosser
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@wdvr wdvr

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@wdvr