Skip to content

Conversation

wnqn1597
Copy link
Contributor

No description provided.

@wnqn1597 wnqn1597 changed the title pnnx convert diag add diag layer and its converter Aug 13, 2023
@wnqn1597 wnqn1597 marked this pull request as ready for review August 13, 2023 12:54
Copy link
Member

@nihui nihui left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

add pnnx ncnn tests
update docs operators.md

@wnqn1597 wnqn1597 requested a review from nihui August 15, 2023 12:30
@codecov-commenter
Copy link

codecov-commenter commented Aug 15, 2023

Codecov Report

Attention: Patch coverage is 95.34884% with 2 lines in your changes missing coverage. Please review.

Project coverage is 94.72%. Comparing base (695f770) to head (9f9040b).
Report is 446 commits behind head on master.

Files with missing lines Patch % Lines
src/layer/diag.cpp 95.34% 2 Missing ⚠️
Additional details and impacted files
@@           Coverage Diff           @@
##           master    #4935   +/-   ##
=======================================
  Coverage   94.72%   94.72%           
=======================================
  Files         770      771    +1     
  Lines      228715   228758   +43     
=======================================
+ Hits       216645   216686   +41     
- Misses      12070    12072    +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@nihui
Copy link
Member

nihui commented Aug 24, 2023

pass_ncnn/torch_diag.cpp doesn't seem to work for some reasons, see pnnx ci errors

@wnqn1597
Copy link
Contributor Author

pass_ncnn/torch_diag.cpp doesn't seem to work for some reasons, see pnnx ci errors

Can I take a look at the logs of pnnx ci?

@nihui
Copy link
Member

nihui commented Aug 28, 2023

pass_ncnn/torch_diag.cpp doesn't seem to work for some reasons, see pnnx ci errors

Can I take a look at the logs of pnnx ci?

click into pnnx ci, login with your github, and you could see the log there

@nihui
Copy link
Member

nihui commented Aug 28, 2023

 446/507 Test #450: test_ncnn_torch_diag ..........................***Failed    5.55 sec
pnnxparam = test_torch_diag.pnnx.param
pnnxbin = test_torch_diag.pnnx.bin
pnnxpy = test_torch_diag_pnnx.py
pnnxonnx = test_torch_diag.pnnx.onnx
ncnnparam = test_torch_diag.ncnn.param
ncnnbin = test_torch_diag.ncnn.bin
ncnnpy = test_torch_diag_ncnn.py
fp16 = 1
optlevel = 2
device = cpu
inputshape = [7]f32,[5,5]f32,[4,8]f32
inputshape2 = 
customop = 
moduleop = 
############# pass_level0

----------------

############# pass_level1
############# pass_level2
############# pass_level3
############# pass_level4
############# pass_level5
############# pass_ncnn
force batch axis 233 for operand 0
force batch axis 233 for operand 4
force batch axis 233 for operand 3
fallback batch axis 233 for operand 1
fallback batch axis 233 for operand 2
fallback batch axis 233 for operand 5
ignore torch.diag torch.diag_2 param diagonal=3
ignore torch.diag torch.diag_1 param diagonal=0
ignore torch.diag torch.diag_0 param diagonal=-1
layer torch.diag not exists or registered
CMake Error at /data/devops/workspace/tools/pnnx/tests/run_test.cmake:5 (message):
  Test failed with return value 'Segmentation fault'

@wnqn1597 wnqn1597 closed this Sep 5, 2023
@wnqn1597 wnqn1597 reopened this Sep 6, 2023
@nihui
Copy link
Member

nihui commented Sep 8, 2023

lgtm

@nihui nihui merged commit c851231 into Tencent:master Sep 8, 2023
@nihui
Copy link
Member

nihui commented Sep 8, 2023

Thanks for your contribution !

@wnqn1597 wnqn1597 deleted the pnnx-convert-diag branch September 9, 2023 09:21
shlyamster pushed a commit to shlyamster/ncnn that referenced this pull request Sep 11, 2023
add diag layer and its converter (Tencent#4935)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants