[GitHub] [tvm] junrushao1994 commented on pull request #10527: [PyTorch][BugFix] PyTorch-TVM Bridge Build Scripts

junrushao1994 commented on pull request #10527: URL: https://github.com/apache/tvm/pull/10527#issuecomment-1063760696 rebased against the latest HEAD...some unfortunate flaky test happened -- This is an automated message from the Apache Git Service. To respond to the message, please log o

[tvm] branch main updated (39bf923 -> a1fb44d)

This is an automated email from the ASF dual-hosted git repository. manupa pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 39bf923 [AOT] Use python temporary directory for AOT tests (#10518) add a1fb44d [microNPU] Add support for TFLite F

[GitHub] [tvm] manupa-arm merged pull request #10345: [microNPU] Add support for TFLite FULLY_CONNECTED

manupa-arm merged pull request #10345: URL: https://github.com/apache/tvm/pull/10345 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsub

[tvm] branch main updated (48793f3 -> 39bf923)

This is an automated email from the ASF dual-hosted git repository. manupa pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 48793f3 Add ONNX LinearRegressor operator support (#10477) add 39bf923 [AOT] Use python temporary directory for AOT

[GitHub] [tvm] manupa-arm commented on pull request #10518: [AOT] Use python temporary directory for AOT tests

manupa-arm commented on pull request #10518: URL: https://github.com/apache/tvm/pull/10518#issuecomment-1063710525 Thanks @lhutton1 . This is merged now! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above t

[GitHub] [tvm] manupa-arm merged pull request #10518: [AOT] Use python temporary directory for AOT tests

manupa-arm merged pull request #10518: URL: https://github.com/apache/tvm/pull/10518 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsub

[GitHub] [tvm] zxybazh commented on pull request #10534: [MetaSchedule] Bug Fix for Relay Integration

zxybazh commented on pull request #10534: URL: https://github.com/apache/tvm/pull/10534#issuecomment-1063679486 Debug info removed and regression tests added. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL ab

[GitHub] [tvm] huajsj opened a new pull request #10560: [Runtime][ThreadPool] Remove a cout log output.

huajsj opened a new pull request #10560: URL: https://github.com/apache/tvm/pull/10560 There is a debug std::cout logic left in thread_pool.cc, remove it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL ab

[GitHub] [tvm] t-vi commented on pull request #10556: Fix TorchScript fallback build

t-vi commented on pull request #10556: URL: https://github.com/apache/tvm/pull/10556#issuecomment-1063666508 Oh oh. Yes. I missed to git add the header - and to enable the tests. Thank you for cleaning this up after me! If I can help beyond what you already have here, let me know.

[tvm] branch last-successful updated (1f60529 -> 48793f3)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git. from 1f60529 [Hexagon] Resolve breakage in test_hexagon/test_cache_read_write (#10520) add 48793f3 Add O

[tvm] branch last-successful updated (060d9d2 -> 1f60529)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git. from 060d9d2 [AOT] Introduce checks for return values from operators (#10424) add 0fa3540 [QNN] Add nn.a

[GitHub] [tvm] qingcd edited a comment on issue #10545: [Bug] TVM model give different results when run multiple times

qingcd edited a comment on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063610234 The gist code: https://gist.github.com/qingcd/6d9d228d92a7b6d09732a6070473a229 I make a simple stupid mistake when copy code on github, just change ' to ''' works. I push the

[GitHub] [tvm] qingcd edited a comment on issue #10545: [Bug] TVM model give different results when run multiple times

qingcd edited a comment on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063610234 The gist code: https://gist.github.com/qingcd/6d9d228d92a7b6d09732a6070473a229 I make a simple stupid mistake when copy code on github, just change "`" to ``` works. I push t

[GitHub] [tvm] qingcd commented on issue #10545: [Bug] TVM model give different results when run multiple times

qingcd commented on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063610234 The gist code: https://gist.github.com/qingcd/6d9d228d92a7b6d09732a6070473a229 I make a simple stupid mistake when copy code on github, just change ` to ``` works. I push the <> but

[GitHub] [tvm] leeexyz commented on pull request #10546: [TE] Bind a rank-0 tensor in `decl_tensor_intrin`

leeexyz commented on pull request #10546: URL: https://github.com/apache/tvm/pull/10546#issuecomment-1063605989 > I'm not sure. Can you explain why "comparison among rank-0 tensor is ambiguous"? I guess [here](https://github.com/apache/tvm/blob/main/python/tvm/te/tensor.py#L89-L95),

[GitHub] [tvm] qingcd opened a new issue #10559: [Bug] TVM model give different results when run multiple times

qingcd opened a new issue #10559: URL: https://github.com/apache/tvm/issues/10559 My pytorch model will have different results when run multiple times with the same input after converting to tvm model. The cuda target fmt is ptx. If the target fmt chage back to cubin, then there is no pro

[GitHub] [tvm] qingcd commented on issue #10545: [Bug] TVM model give different results when run multiple times

qingcd commented on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063596070 Thanks for your help, I'll try! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the spec

[GitHub] [tvm] masahi commented on issue #10545: [Bug] TVM model give different results when run multiple times

masahi commented on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063594668 You can use https://gist.github.com rather than pasting your code on github -- This is an automated message from the Apache Git Service. To respond to the message, please log on to Gi

[GitHub] [tvm] qingcd edited a comment on issue #10545: [Bug] TVM model give different results when run multiple times

qingcd edited a comment on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063585897 Thanks for your attention. There may be some problems with the syntax of github bug report, the code display is very wired, I spend hours but can not fix it, what's more, gith

[GitHub] [tvm] qingcd edited a comment on issue #10545: [Bug] TVM model give different results when run multiple times

qingcd edited a comment on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063585897 Thanks for your attention. This example can work, just copy and run. There may be some problems with the syntax of github bug report, the code display is very wired, I spen

[GitHub] [tvm] Hzfengsy commented on pull request #10546: [TE] Bind a rank-0 tensor in `decl_tensor_intrin`

Hzfengsy commented on pull request #10546: URL: https://github.com/apache/tvm/pull/10546#issuecomment-1063587373 I'm not sure. Can you explain why "comparison among rank-0 tensor is ambiguous"? -- This is an automated message from the Apache Git Service. To respond to the message, please

[GitHub] [tvm] qingcd commented on issue #10545: [Bug]

qingcd commented on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063585897 Thanks for your attention. This example can work, just copy and run. There may be some problems with the syntax of github bug report, the code display is very wired, I spend hours

[GitHub] [tvm] Hzfengsy commented on pull request #10504: complete pytorch grid_sample

Hzfengsy commented on pull request #10504: URL: https://github.com/apache/tvm/pull/10504#issuecomment-1063579934 Please rebase it and then we can review -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to

[GitHub] [tvm] masahi closed issue #10545: [Bug]

masahi closed issue #10545: URL: https://github.com/apache/tvm/issues/10545 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tv

[GitHub] [tvm] masahi commented on issue #10545: [Bug]

masahi commented on issue #10545: URL: https://github.com/apache/tvm/issues/10545#issuecomment-1063574994 Make sure you get more basic examples working first. And post a complete working repro script, what you showed is not really executable. It's probably a general usage question n

[tvm] branch main updated (1f60529 -> 48793f3)

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 1f60529 [Hexagon] Resolve breakage in test_hexagon/test_cache_read_write (#10520) add 48793f3 Add ONNX LinearReg

[GitHub] [tvm] comaniac commented on pull request #10477: [Frontend][ONNX] Add LinearRegressor operator support

comaniac commented on pull request #10477: URL: https://github.com/apache/tvm/pull/10477#issuecomment-1063566586 Thanks @zxy844288792 @huajsj -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the

[GitHub] [tvm] comaniac merged pull request #10477: [Frontend][ONNX] Add LinearRegressor operator support

comaniac merged pull request #10477: URL: https://github.com/apache/tvm/pull/10477 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubsc

[GitHub] [tvm] comaniac commented on a change in pull request #10477: [Frontend][ONNX] Add LinearRegressor operator support

comaniac commented on a change in pull request #10477: URL: https://github.com/apache/tvm/pull/10477#discussion_r823260409 ## File path: python/tvm/relay/frontend/onnx.py ## @@ -3444,6 +3444,35 @@ def body_fn(*loop_inputs): return outputs +class LinearRegressor(Onn

[tvm] branch last-successful updated (f9d3918 -> 060d9d2)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git. from f9d3918 [LLVM][TIR] Propagate variable names to parameters. (#10514) add 060d9d2 [AOT] Introduce che

[GitHub] [tvm] masahi commented on pull request #10556: Fix TorchScript fallback build

masahi commented on pull request #10556: URL: https://github.com/apache/tvm/pull/10556#issuecomment-1063557540 please fix the lint issue. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the spec

[tvm] branch main updated (0fa3540 -> 1f60529)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 0fa3540 [QNN] Add nn.adaptive_avg_pool1d to FQ2I (#10541) add 1f60529 [Hexagon] Resolve breakage in test_hexagon/t

[GitHub] [tvm] masahi merged pull request #10520: [Hexagon] Resolve breakage in test_hexagon/test_cache_read_write

masahi merged pull request #10520: URL: https://github.com/apache/tvm/pull/10520 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr.

[tvm] branch main updated (060d9d2 -> 0fa3540)

This is an automated email from the ASF dual-hosted git repository. mbrookhart pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 060d9d2 [AOT] Introduce checks for return values from operators (#10424) add 0fa3540 [QNN] Add nn.adaptive_avg

[GitHub] [tvm] adstraw commented on pull request #10520: [Hexagon] Resolve breakage in test_hexagon/test_cache_read_write

adstraw commented on pull request #10520: URL: https://github.com/apache/tvm/pull/10520#issuecomment-1063539029 Still LGTM with latest changes and I can confirm the test in questions (test_cache_read_write.py) is passing locally. -- This is an automated message from the Apache Git Servic

[GitHub] [tvm] mbrookhart commented on pull request #10541: [QNN] Add nn.adaptive_avg_pool1d to FQ2I

mbrookhart commented on pull request #10541: URL: https://github.com/apache/tvm/pull/10541#issuecomment-1063537125 Thanks @MargaretQian @anwang2009 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go

[GitHub] [tvm] mbrookhart merged pull request #10541: [QNN] Add nn.adaptive_avg_pool1d to FQ2I

mbrookhart merged pull request #10541: URL: https://github.com/apache/tvm/pull/10541 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsub

[GitHub] [tvm] junrushao1994 commented on pull request #10557: [TIR] compact buffer region

junrushao1994 commented on pull request #10557: URL: https://github.com/apache/tvm/pull/10557#issuecomment-1063523182 Thanks for the fix! It's essential for AutoTIR performance!! Please add a regression test :-) -- This is an automated message from the Apache Git Service. To respond to t

[GitHub] [tvm] spectrometerHBH opened a new pull request #10557: [TIR] compact buffer region

spectrometerHBH opened a new pull request #10557: URL: https://github.com/apache/tvm/pull/10557 The compacted buffer's shape should be at most the same as the original shape. cc @junrushao1994 -- This is an automated message from the Apache Git Service. To respond to the message, ple

[GitHub] [tvm] junrushao1994 commented on pull request #10551: [TVMScript] Add intrinsic to look up llvm intrinsic id

junrushao1994 commented on pull request #10551: URL: https://github.com/apache/tvm/pull/10551#issuecomment-1063518282 Looks like the CI failed due to: ``` ModuleNotFoundError: No module named 'tvm.script.target' ``` -- This is an automated message from the Apache Git Service.

[GitHub] [tvm] areusch commented on a change in pull request #10539: [ci] Build GPU libraries on CPU nodes

areusch commented on a change in pull request #10539: URL: https://github.com/apache/tvm/pull/10539#discussion_r823205516 ## File path: tests/scripts/task_build.py ## @@ -32,6 +32,7 @@ parser.add_argument("--sccache-bucket", required=False, help="sccache bucket name")

[GitHub] [tvm] areusch commented on a change in pull request #10537: [ci] Remove commit check on ci skipping logic

areusch commented on a change in pull request #10537: URL: https://github.com/apache/tvm/pull/10537#discussion_r823202330 ## File path: tests/scripts/git_skip_ci.py ## @@ -49,13 +49,7 @@ def check_pr_title(): print("pr title:", title) return title.startswith("

[GitHub] [tvm] areusch commented on pull request #10257: [ci] Move pip dependencies to docker images, add ninja / shellcheck

areusch commented on pull request #10257: URL: https://github.com/apache/tvm/pull/10257#issuecomment-1063495481 i think maybe we should change task_ci_setup.sh to not install those packages if they already exist. i can't remember exactly what happened here, but there was a CI problem befor

[GitHub] [tvm] huajsj commented on a change in pull request #10477: [Frontend][ONNX] Add LinearRegressor operator support

huajsj commented on a change in pull request #10477: URL: https://github.com/apache/tvm/pull/10477#discussion_r823186719 ## File path: python/tvm/relay/frontend/onnx.py ## @@ -3444,6 +3444,35 @@ def body_fn(*loop_inputs): return outputs +class LinearRegressor(OnnxO

[GitHub] [tvm] driazati opened a new pull request #10556: Fix TorchScript fallback build

driazati opened a new pull request #10556: URL: https://github.com/apache/tvm/pull/10556 This was missing a header `libtorch_runtime.h`. The test in `test_libtorch_ops.py` is also currently being skipped in CI since `torch` isn't available but that's left for a follow up cc @t-vi @m

[GitHub] [tvm] huajsj commented on a change in pull request #10530: [Driver] Remove duplicate PreProcessModuleForBuild

huajsj commented on a change in pull request #10530: URL: https://github.com/apache/tvm/pull/10530#discussion_r823179659 ## File path: python/tvm/driver/build_module.py ## @@ -275,7 +275,7 @@ def build( annotated_mods, target_host = Target.check_and_update_host_consist(

[tvm] branch last-successful updated (fe33ed6 -> f9d3918)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git. from fe33ed6 RFC: initial stab at TorchScript fallback (#7401) add f9d3918 [LLVM][TIR] Propagate variable

[GitHub] [tvm] masahi edited a comment on issue #10554: [Flakey] tests/python/frontend/pytorch/test_forward.py::test_forward_roi_align

masahi edited a comment on issue #10554: URL: https://github.com/apache/tvm/issues/10554#issuecomment-1063461749 hmm, I don't recall flaky-ness from this test. As long as we use random inputs, there is always non-zero prob of failing, I'd say. -- This is an automated message from the Apa

[GitHub] [tvm] masahi commented on issue #10554: [Flakey] tests/python/frontend/pytorch/test_forward.py::test_forward_roi_align

masahi commented on issue #10554: URL: https://github.com/apache/tvm/issues/10554#issuecomment-1063461749 hmm, I don't recall flaky-ness from this test. As long as we use the random input, there is always non-zero prob of failing, I'd say. -- This is an automated message from the Apache

[tvm] branch main updated (f9d3918 -> 060d9d2)

This is an automated email from the ASF dual-hosted git repository. manupa pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from f9d3918 [LLVM][TIR] Propagate variable names to parameters. (#10514) add 060d9d2 [AOT] Introduce checks for return

[GitHub] [tvm] manupa-arm merged pull request #10424: [AOT] Introduce checks for return values from operators

manupa-arm merged pull request #10424: URL: https://github.com/apache/tvm/pull/10424 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsub

[GitHub] [tvm] manupa-arm commented on pull request #10424: [AOT] Introduce checks for return values from operators

manupa-arm commented on pull request #10424: URL: https://github.com/apache/tvm/pull/10424#issuecomment-1063429313 Thanks! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment.

[tvm] branch last-successful updated (ffd5f70 -> fe33ed6)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git. from ffd5f70 Revert "[TE][Fix] Comparison of the output tensor (#9829)" (#10540) add fe33ed6 RFC: initia

[GitHub] [tvm] mehrdadh opened a new pull request #10555: [microTVM] Enable micro tvmc tutorial testing in CI

mehrdadh opened a new pull request #10555: URL: https://github.com/apache/tvm/pull/10555 This PR changes micro_tvmc.py tutorial to a bash script tutorial. Also it adds a python script which converts a bash script tutorial to python format for doc generation. I have added the script

[GitHub] [tvm] tkonolige commented on pull request #10532: [FIX] Remove hexagon files built when hexagon is disabled

tkonolige commented on pull request #10532: URL: https://github.com/apache/tvm/pull/10532#issuecomment-1063410990 @csullivan 's solution seems to satisfy everyone's constraints here. It won't require setting global flags, and it will correctly disable log customization on regular targets.

[GitHub] [tvm] driazati commented on pull request #10323: Clean up issue templates

driazati commented on pull request #10323: URL: https://github.com/apache/tvm/pull/10323#issuecomment-1063405600 Moving to a draft until we can align better on https://discuss.tvm.apache.org/t/consolidate-development-activity-on-github/12219 -- This is an automated message from the Apach

[GitHub] [tvm] github-actions[bot] commented on pull request #10323: Clean up issue templates

github-actions[bot] commented on pull request #10323: URL: https://github.com/apache/tvm/pull/10323#issuecomment-1063404673 It has been a while since this PR was updated, @driazati @areusch @denise-k @hpanda-naut @areusch please leave a review or address the outstanding comments -- This

[GitHub] [tvm] driazati commented on issue #10554: [Flakey] tests/python/frontend/pytorch/test_forward.py::test_forward_roi_align

driazati commented on issue #10554: URL: https://github.com/apache/tvm/issues/10554#issuecomment-1063397642 cc @masahi any ideas why the numerics wouldnt be stable across tests / if we can just increase the tolerance a bit on this test? -- This is an automated message from the Apache Git

[GitHub] [tvm] tkonolige opened a new issue #10554: [Flakey] tests/python/frontend/pytorch/test_forward.py::test_forward_roi_align

tkonolige opened a new issue #10554: URL: https://github.com/apache/tvm/issues/10554 tests/python/frontend/pytorch/test_forward.py::test_forward_roi_align appears to be flakey: https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-10550/1/pipeline. @areusch @driazati --

[GitHub] [tvm] junrushao1994 commented on pull request #10546: [TE] Bind a rank-0 tensor in `decl_tensor_intrin`

junrushao1994 commented on pull request #10546: URL: https://github.com/apache/tvm/pull/10546#issuecomment-1063381296 CC: @Hzfengsy -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific

[GitHub] [tvm] junrushao1994 commented on pull request #10551: [TVMScript] Add intrinsic to look up llvm intrinsic id

junrushao1994 commented on pull request #10551: URL: https://github.com/apache/tvm/pull/10551#issuecomment-1063380345 looks like there is some issue with the CI -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [tvm] Lunderberg opened a new pull request #10553: [Bugfix] Simultaneous layout transform and axis separators.

Lunderberg opened a new pull request #10553: URL: https://github.com/apache/tvm/pull/10553 Previously, SchedulePostProcToPrimFunc would first generate the map from buffer object to layout transformation, then would update buffers with the axis separators. However, it failed to replace the

[tvm] branch main updated (fe33ed6 -> f9d3918)

This is an automated email from the ASF dual-hosted git repository. kparzysz pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from fe33ed6 RFC: initial stab at TorchScript fallback (#7401) add f9d3918 [LLVM][TIR] Propagate variable names to par

[GitHub] [tvm] kparzysz-quic merged pull request #10514: [LLVM][TIR] Propagate variable names to parameters.

kparzysz-quic merged pull request #10514: URL: https://github.com/apache/tvm/pull/10514 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-un

[GitHub] [tvm] t-vi commented on pull request #7401: RFC: initial stab at TorchScript fallback

t-vi commented on pull request #7401: URL: https://github.com/apache/tvm/pull/7401#issuecomment-1063277207 Thank you @masahi ! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific commen

[tvm] branch main updated (ffd5f70 -> fe33ed6)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from ffd5f70 Revert "[TE][Fix] Comparison of the output tensor (#9829)" (#10540) add fe33ed6 RFC: initial stab at Torch

[GitHub] [tvm] masahi merged pull request #7401: RFC: initial stab at TorchScript fallback

masahi merged pull request #7401: URL: https://github.com/apache/tvm/pull/7401 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...

[GitHub] [tvm] Lunderberg commented on pull request #10520: [Hexagon] Resolve breakage in test_hexagon/test_cache_read_write

Lunderberg commented on pull request #10520: URL: https://github.com/apache/tvm/pull/10520#issuecomment-1063260297 After a discussion with @adstraw , updated the fix here to remove the `StackSizeChecker` entirely. Instead of having two cases that must be kept in sync, it instead runs the

[GitHub] [tvm] tkonolige opened a new pull request #10552: [FIX,HEXAGON] Gitignore generated Hexagon files

tkonolige opened a new pull request #10552: URL: https://github.com/apache/tvm/pull/10552 Ignore these files so they are not accidentally committed. @kparzysz-quic @adstraw @csullivan -- This is an automated message from the Apache Git Service. To respond to the message, please lo

[GitHub] [tvm] vinx13 opened a new pull request #10551: [TVMScript] Add intrinsic to look up llvm intrinsic id

vinx13 opened a new pull request #10551: URL: https://github.com/apache/tvm/pull/10551 cc @masahi @junrushao1994 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubs

[GitHub] [tvm] tkonolige opened a new pull request #10550: [TVMSCRIPT] Add type definition for preflattened_buffer

tkonolige opened a new pull request #10550: URL: https://github.com/apache/tvm/pull/10550 Seems like this type annotation was missed with Eric's big PR. @junrushao1994 @Lunderberg -- This is an automated message from the Apache Git Service. To respond to the message, please log on

[GitHub] [tvm] tkonolige commented on a change in pull request #10455: [AUTO_SCHEDULER] Add feature extraction directly from PrimFunc

tkonolige commented on a change in pull request #10455: URL: https://github.com/apache/tvm/pull/10455#discussion_r822959339 ## File path: python/tvm/auto_scheduler/feature.py ## @@ -252,3 +253,78 @@ def get_per_store_feature_names(max_n_bufs: Optional[int] = None) -> List[str]

[GitHub] [tvm] tkonolige commented on a change in pull request #10455: [AUTO_SCHEDULER] Add feature extraction directly from PrimFunc

tkonolige commented on a change in pull request #10455: URL: https://github.com/apache/tvm/pull/10455#discussion_r822958745 ## File path: src/auto_scheduler/feature.cc ## @@ -659,38 +697,86 @@ class PerStoreFeatureExtractor : public StmtExprVisitor { } } + void Visit

[GitHub] [tvm] anwang2009 commented on pull request #10541: [QNN] Add nn.adaptive_avg_pool1d to FQ2I

anwang2009 commented on pull request #10541: URL: https://github.com/apache/tvm/pull/10541#issuecomment-1063227392 LGTM! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To

[GitHub] [tvm] shukun-ziqiangxu closed pull request #10504: complete pytorch grid_sample

shukun-ziqiangxu closed pull request #10504: URL: https://github.com/apache/tvm/pull/10504 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits

[GitHub] [tvm] adstraw commented on pull request #10532: [FIX] Remove hexagon files built when hexagon is disabled

adstraw commented on pull request #10532: URL: https://github.com/apache/tvm/pull/10532#issuecomment-1063157137 > If @adstraw prefers to keep these checks would it be acceptable to only enable custom logging when those sources are built for Hexagon? I prefer to keep the pre-commit ch

[GitHub] [tvm] grant-arm commented on pull request #10549: [microNPU] Eliminate some compiler warnings for microNPU demo app

grant-arm commented on pull request #10549: URL: https://github.com/apache/tvm/pull/10549#issuecomment-1063127859 @Mousius @manupa-arm @areusch -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to t

[GitHub] [tvm] grant-arm opened a new pull request #10549: [microNPU] Eliminate some compiler warnings for microNPU demo app

grant-arm opened a new pull request #10549: URL: https://github.com/apache/tvm/pull/10549 * Eliminates all "control reaches end of non-void function" warnings * Eliminates some "implicit declaration of function" warnings * Includes tvm_ethosu_runtime.c and tvm_ethosu_runtime.h in ML

[GitHub] [tvm] kparzysz-quic commented on pull request #10532: [FIX] Remove hexagon files built when hexagon is disabled

kparzysz-quic commented on pull request #10532: URL: https://github.com/apache/tvm/pull/10532#issuecomment-1063121256 We should have this functionality tested in CI, but it should be done on Hexagon. We should figure out how we can run these tests, or their equivalents on simulator. The

[GitHub] [tvm] ashutosh-arm opened a new pull request #10548: [CMSIS-NN] Include clip in the qnn binary op patterns

ashutosh-arm opened a new pull request #10548: URL: https://github.com/apache/tvm/pull/10548 Support clip op in qnn binary op patterns for CMSIS-NN -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go t

[GitHub] [tvm] csullivan commented on pull request #10532: [FIX] Remove hexagon files built when hexagon is disabled

csullivan commented on pull request #10532: URL: https://github.com/apache/tvm/pull/10532#issuecomment-1063110308 @adstraw's intention with adding these tests was to allow for unit testing of some of the C++ classes that would run pre-commit. Of course in production they run on Hexagon whe

[GitHub] [tvm] kparzysz-quic commented on pull request #10465: [tir.constant] bugfix 10390

kparzysz-quic commented on pull request #10465: URL: https://github.com/apache/tvm/pull/10465#issuecomment-1063106533 Could you rebase this PR? Is this the final version? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and u

[GitHub] [tvm] manupa-arm commented on issue #10487: [Flaky Test] ethosu `test_networks_without_usmp` and `test_mean`

manupa-arm commented on issue #10487: URL: https://github.com/apache/tvm/issues/10487#issuecomment-1063105756 @driazati, > Sure we can do that moving forward, but if a test fails on more than 1 parameterized version I think we should still stick with disabling the whole test to redu

[GitHub] [tvm] d-smirnov opened a new pull request #10547: [USMP] HillClimb stability patch

d-smirnov opened a new pull request #10547: URL: https://github.com/apache/tvm/pull/10547 This patch increases stability of the hill climb allocation algorithm @manupa-arm -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub a

[GitHub] [tvm] ekalda commented on a change in pull request #10508: [microNPU] Improve cascader memory transfer estimates

ekalda commented on a change in pull request #10508: URL: https://github.com/apache/tvm/pull/10508#discussion_r822724373 ## File path: python/tvm/contrib/ethosu/cascader/tensor_config.py ## @@ -58,9 +58,25 @@ class MemoryRegion(Object): """ -def __init__(self, name

[GitHub] [tvm] Mousius commented on a change in pull request #10424: [AOT] Introduce checks for return values from operators

Mousius commented on a change in pull request #10424: URL: https://github.com/apache/tvm/pull/10424#discussion_r822821490 ## File path: src/relay/backend/aot_executor_codegen.cc ## @@ -333,6 +333,19 @@ class AOTExecutorCodegen : public MixedModeVisitor { args->insert(args-

[GitHub] [tvm] manupa-arm commented on a change in pull request #10424: [AOT] Introduce checks for return values from operators

manupa-arm commented on a change in pull request #10424: URL: https://github.com/apache/tvm/pull/10424#discussion_r822816091 ## File path: src/relay/backend/aot_executor_codegen.cc ## @@ -333,6 +333,19 @@ class AOTExecutorCodegen : public MixedModeVisitor { args->insert(ar

[GitHub] [tvm] Lunderberg commented on pull request #10514: [LLVM][TIR] Propagate variable names to parameters.

Lunderberg commented on pull request #10514: URL: https://github.com/apache/tvm/pull/10514#issuecomment-1063028014 Rebase onto main following #10535 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go

[GitHub] [tvm] Lunderberg commented on pull request #10516: [Relay][TE] Use Relay parameter name to generated TE tensor name

Lunderberg commented on pull request #10516: URL: https://github.com/apache/tvm/pull/10516#issuecomment-1063024066 Rebase onto main following #10535 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go

[GitHub] [tvm] Lunderberg removed a comment on pull request #10514: [LLVM][TIR] Propagate variable names to parameters.

Lunderberg removed a comment on pull request #10514: URL: https://github.com/apache/tvm/pull/10514#issuecomment-1063023402 Rebase onto main following #10535 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL abov

[GitHub] [tvm] Lunderberg commented on pull request #10514: [LLVM][TIR] Propagate variable names to parameters.

Lunderberg commented on pull request #10514: URL: https://github.com/apache/tvm/pull/10514#issuecomment-1063023402 Rebase onto main following #10535 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go

[GitHub] [tvm] Lunderberg commented on pull request #10530: [Driver] Remove duplicate PreProcessModuleForBuild

Lunderberg commented on pull request #10530: URL: https://github.com/apache/tvm/pull/10530#issuecomment-1063010060 Rebased onto main following #10535. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to g

[GitHub] [tvm] yuanfz98 closed pull request #10544: [TIR] Semantic extension for CSE Pass

yuanfz98 closed pull request #10544: URL: https://github.com/apache/tvm/pull/10544 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubsc

[GitHub] [tvm] yuanfz98 commented on a change in pull request #10544: [TIR] Semantic extension for CSE Pass

yuanfz98 commented on a change in pull request #10544: URL: https://github.com/apache/tvm/pull/10544#discussion_r822739004 ## File path: src/tir/transforms/common_subexpr_elim_tools.cc ## @@ -727,7 +728,10 @@ bool EquivalentTerms(const PrimExpr& a, const PrimExpr& b) { // F

[tvm] 01/01: [CI] Enable TOPI tests in ci_arm

This is an automated email from the ASF dual-hosted git repository. mousius pushed a commit to branch more-topi in repository https://gitbox.apache.org/repos/asf/tvm.git commit 46d8e5001feb6f5bb6a6f28cd18dbc9de646d841 Author: Chris Sidebottom AuthorDate: Wed Mar 9 10:57:08 2022 + [CI] E

[tvm] branch more-topi updated (ebc086a -> 46d8e50)

This is an automated email from the ASF dual-hosted git repository. mousius pushed a change to branch more-topi in repository https://gitbox.apache.org/repos/asf/tvm.git. discard ebc086a [CI] Enable TOPI tests in ci_arm new 46d8e50 [CI] Enable TOPI tests in ci_arm This update added new

[GitHub] [tvm] tqchen commented on a change in pull request #10544: [TIR] Semantic extension for CSE Pass

tqchen commented on a change in pull request #10544: URL: https://github.com/apache/tvm/pull/10544#discussion_r822686615 ## File path: src/tir/transforms/common_subexpr_elim_tools.cc ## @@ -727,7 +728,10 @@ bool EquivalentTerms(const PrimExpr& a, const PrimExpr& b) { // For

[tvm] branch last-successful updated (a8a2574 -> ffd5f70)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch last-successful in repository https://gitbox.apache.org/repos/asf/tvm.git. from a8a2574 [skip ci][Bugfix] Allow constant folding of 0U - 0U (#10535) add ffd5f70 Revert "[TE][Fix] C

[GitHub] [tvm] dchauhan-arm commented on a change in pull request #10345: [microNPU] Add support for TFLite FULLY_CONNECTED

dchauhan-arm commented on a change in pull request #10345: URL: https://github.com/apache/tvm/pull/10345#discussion_r822613407 ## File path: tests/python/contrib/test_ethosu/test_legalize.py ## @@ -2421,5 +2421,108 @@ def verify(ext_func): verify(mod["tvmgen_default_ethos_

[GitHub] [tvm] leeexyz opened a new pull request #10546: [TE] Bind a rank-0 tensor in `decl_tensor_intrin`

leeexyz opened a new pull request #10546: URL: https://github.com/apache/tvm/pull/10546 Hi All, This PR addressed issue 1 in https://github.com/apache/tvm/pull/9829. This can solve the following error. ```bash ~/project/apache/tvm/tests/python/unittest$ pytest test_te_schedule.

  1   2   >