This is an automated email from the ASF dual-hosted git repository.
junrushao pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from ce8ac3e784 [TIR] Allow missing TensorIntrin in registry lookup (#12875)
add 9ce95a9abe [TIR] Fix wmma index in
junrushao merged PR #12879:
URL: https://github.com/apache/tvm/pull/12879
--
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:
This is an automated email from the ASF dual-hosted git repository.
junrushao pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 0b074d8f06 Fix clang warnings (#12876)
add ce8ac3e784 [TIR] Allow missing TensorIntrin in registry lookup (#12875)
junrushao merged PR #12875:
URL: https://github.com/apache/tvm/pull/12875
--
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:
DzAvril opened a new pull request, #12881:
URL: https://github.com/apache/tvm/pull/12881
```c++
Module GraphExecutorFactory::ExecutorCreate(const std::vector& devs)
{
auto exec = make_object();
exec->Init(this->graph_json_, this->imports_[0], devs, PackedFunc());
// set
github-actions[bot] commented on PR #12880:
URL: https://github.com/apache/tvm/pull/12880#issuecomment-1255742179
Thanks for contributing to TVM! Please refer to the contributing guidelines
https://tvm.apache.org/docs/contribute/ for useful information and tips. Please
request code
areusch commented on PR #12880:
URL: https://github.com/apache/tvm/pull/12880#issuecomment-1255740385
NOTE: sending this through CI to see what I broke, then will add tests.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
areusch opened a new pull request, #12880:
URL: https://github.com/apache/tvm/pull/12880
* Fixes cases where a USMP buffer is used with tensorization intrinsics. In
these cases, the buffer data var is referenced directly in the call_extern args.
Before this patch,
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 4e783a6087 [TOPI] Add layer norm operator (#12864)
add 0b074d8f06 Fix clang warnings (#12876)
No new
This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch nightly-docker-update
in repository https://gitbox.apache.org/repos/asf/tvm.git
omit 4671a5743b [ci][docker] Nightly Docker image update
add 39f71ae288 [frontend][pytorch] Add a new
vinx13 opened a new pull request, #12879:
URL: https://github.com/apache/tvm/pull/12879
There are some corner cases `get_wmma_fragment_index` returns wrong result
when the size of shared memory is large. In practice, such case usually won't
happen because the size of the shared memory is
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 195ae72b5c [TOPI] Fix dtype legalize logic for CPU dot product
instruction (#12865)
add 86f9580498
junrushao merged PR #12876:
URL: https://github.com/apache/tvm/pull/12876
--
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:
This is an automated email from the ASF dual-hosted git repository.
junrushao pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 4e783a6087 [TOPI] Add layer norm operator (#12864)
add 0b074d8f06 Fix clang warnings (#12876)
No new revisions
tmoreau89 commented on PR #12877:
URL: https://github.com/apache/tvm/pull/12877#issuecomment-1255566577
CC @Lunderberg @areusch @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
mehrdadh commented on code in PR #12818:
URL: https://github.com/apache/tvm/pull/12818#discussion_r978098988
##
apps/microtvm/arduino/template_project/microtvm_api_server.py:
##
@@ -313,7 +297,90 @@ def _find_modified_include_path(self, project_dir,
file_path, include_path):
mehrdadh commented on code in PR #12818:
URL: https://github.com/apache/tvm/pull/12818#discussion_r978098147
##
apps/microtvm/arduino/template_project/microtvm_api_server.py:
##
@@ -313,7 +297,90 @@ def _find_modified_include_path(self, project_dir,
file_path, include_path):
This is an automated email from the ASF dual-hosted git repository.
areusch pushed a commit to branch areusch/fix-retry-main
in repository https://gitbox.apache.org/repos/asf/tvm.git
commit efbff7ad7fcc34d3af00857381511ce003939494
Author: Andrew Reusch
AuthorDate: Thu Sep 22 14:23:25 2022 -0700
This is an automated email from the ASF dual-hosted git repository.
areusch pushed a change to branch areusch/fix-retry-main
in repository https://gitbox.apache.org/repos/asf/tvm.git
at efbff7ad7f git_init before Deploy to get retry.sh.
This branch includes the following new commits:
areusch opened a new issue, #12878:
URL: https://github.com/apache/tvm/issues/12878
See
https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/main/4345/pipeline
Looks like #12773 made it necessary to checkout the repo into the post-build
steps. PR to follow.
--
This is an
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 fe75f00991 [AutoTVM] Introducing multi_filter into ConfigSpace autotvm
(#12545)
add 195ae72b5c [TOPI]
slyubomirsky commented on PR #12682:
URL: https://github.com/apache/tvm/pull/12682#issuecomment-1255542976
Hm, some of the builds still use different versions of CMake. Any pointers
on where the CMake versions for those are defined? I would be glad to make the
changes myself.
--
This is
alanmacd commented on code in PR #12851:
URL: https://github.com/apache/tvm/pull/12851#discussion_r978078873
##
include/tvm/runtime/crt/rpc_common/write_stream.h:
##
@@ -29,6 +29,7 @@
#include
#include
#include
+#include
Review Comment:
@mkatanbaf @areusch I moved
kparzysz-quic opened a new pull request, #12877:
URL: https://github.com/apache/tvm/pull/12877
For conversions between `_Float16` and `float`, LLVM uses runtime functions
`__extendhfsf2` and `__truncsfhf2`. On X86 up until version 14, LLVM used
`uint16_t` for representing `_Float16`.
vinx13 merged PR #12864:
URL: https://github.com/apache/tvm/pull/12864
--
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:
This is an automated email from the ASF dual-hosted git repository.
wuwei pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 86f9580498 [Relay] Fix handling of TransfromLayout in TE compiler
cache (#12874)
add 4e783a6087 [TOPI] Add layer norm
vinx13 commented on code in PR #12864:
URL: https://github.com/apache/tvm/pull/12864#discussion_r978051047
##
python/tvm/topi/nn/__init__.py:
##
@@ -38,6 +38,7 @@
from .bnn import *
from .qnn import *
from .upsampling import *
+from .layer_norm import layer_norm
Review
vinx13 merged PR #12874:
URL: https://github.com/apache/tvm/pull/12874
--
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:
This is an automated email from the ASF dual-hosted git repository.
wuwei pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 195ae72b5c [TOPI] Fix dtype legalize logic for CPU dot product
instruction (#12865)
add 86f9580498 [Relay] Fix
areusch commented on code in PR #12818:
URL: https://github.com/apache/tvm/pull/12818#discussion_r978035114
##
apps/microtvm/arduino/template_project/microtvm_api_server.py:
##
@@ -313,7 +297,90 @@ def _find_modified_include_path(self, project_dir,
file_path, include_path):
zxybazh commented on PR #12866:
URL: https://github.com/apache/tvm/pull/12866#issuecomment-1255484956
Hi YJ, I think this applies to all the ipython kernel based interface, yet I
only tested on jupyter notebook. On the other hand we have progress as
percentage available now so I'm not
mbrookhart commented on issue #12707:
URL: https://github.com/apache/tvm/issues/12707#issuecomment-1255479227
Hi @honghuichao. I'm way behind on this, I think I see a fix.
I did special handling for quantize and dequantize in the
shingjan commented on issue #12651:
URL: https://github.com/apache/tvm/issues/12651#issuecomment-1255424901
+1
--
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
junrushao commented on PR #12874:
URL: https://github.com/apache/tvm/pull/12874#issuecomment-1255420160
Agreed. Perhaps on this particular case the test is not required
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use
junrushao commented on PR #12875:
URL: https://github.com/apache/tvm/pull/12875#issuecomment-1255419263
Got it! Can we add a simple test where the intrin name is missing?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and
vinx13 commented on PR #12874:
URL: https://github.com/apache/tvm/pull/12874#issuecomment-1255419272
it's hard to test without running e2e. alternatively we can get lookup inst
trait registey to get num attrs
--
This is an automated message from the Apache Git Service.
To respond to the
vinx13 commented on PR #12875:
URL: https://github.com/apache/tvm/pull/12875#issuecomment-1255417572
One use case is check if tensor interns exists, it will allow further
automation like generate tensor intrins if missing
--
This is an automated message from the Apache Git Service.
To
junrushao commented on code in PR #12875:
URL: https://github.com/apache/tvm/pull/12875#discussion_r977974861
##
include/tvm/tir/function.h:
##
@@ -233,10 +233,13 @@ class TensorIntrin : public ObjectRef {
/*!
* \brief Look up TensorIntrin by name. Raises an exception if
junrushao commented on code in PR #12875:
URL: https://github.com/apache/tvm/pull/12875#discussion_r977975093
##
python/tvm/tir/function.py:
##
@@ -228,20 +228,24 @@ def register(name: str, desc: PrimFunc, impl: PrimFunc):
return _ffi_api.TensorIntrinRegister(name,
junrushao commented on PR #12875:
URL: https://github.com/apache/tvm/pull/12875#issuecomment-1255408219
Would you like to provide some context BTW? like, in which cases do we allow
it missing?
--
This is an automated message from the Apache Git Service.
To respond to the message, please
vinx13 opened a new pull request, #12876:
URL: https://github.com/apache/tvm/pull/12876
cc @junrushao
--
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,
This is an automated email from the ASF dual-hosted git repository.
tkonolige pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git
from fe75f00991 [AutoTVM] Introducing multi_filter into ConfigSpace autotvm
(#12545)
add 195ae72b5c [TOPI] Fix dtype
tkonolige merged PR #12865:
URL: https://github.com/apache/tvm/pull/12865
--
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:
vinx13 opened a new pull request, #12875:
URL: https://github.com/apache/tvm/pull/12875
Added an option to allow missing tensor intrin.
cc @junrushao @spectrometerHBH
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
shingjan commented on PR #12866:
URL: https://github.com/apache/tvm/pull/12866#issuecomment-1255370680
This seems super useful. Do we want to extends this functionality to
non-jupyter notebook? Essentially what I have in mind is that we can update the
table & process bar and omit all other
zxybazh commented on PR #12866:
URL: https://github.com/apache/tvm/pull/12866#issuecomment-1255333056
@tvm-bot rerun
--
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
This is an automated email from the ASF dual-hosted git repository.
andrewzhaoluo pushed a commit to branch v0.10.0
in repository https://gitbox.apache.org/repos/asf/tvm.git
The following commit(s) were added to refs/heads/v0.10.0 by this push:
new 5328c6bf77 [Hybrid] Fix sys version check
vinx13 opened a new pull request, #12874:
URL: https://github.com/apache/tvm/pull/12874
Because of recent updates `TransformLayout` now has 4 attrs.
https://github.com/apache/tvm/blob/main/src/tir/schedule/primitive/layout_transformation.cc#L1348
cc @junrushao
--
This is an
alanmacd commented on PR #12851:
URL: https://github.com/apache/tvm/pull/12851#issuecomment-1255298527
@tvm-bot rerun
--
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
tkonolige commented on code in PR #12865:
URL: https://github.com/apache/tvm/pull/12865#discussion_r977853794
##
tests/python/relay/test_op_level2.py:
##
@@ -2137,5 +2137,80 @@ def get_subgraph(dtype):
np.testing.assert_allclose(out, ref, rtol=1e-5, atol=1e-5)
areusch commented on code in PR #12691:
URL: https://github.com/apache/tvm/pull/12691#discussion_r977856513
##
src/te/schedule/schedule_lang.cc:
##
@@ -200,7 +200,7 @@ Stage& Stage::env_threads(Array threads) {
ICHECK_EQ(self->env_threads.size(), 0U) << "Already set
tkonolige commented on code in PR #12865:
URL: https://github.com/apache/tvm/pull/12865#discussion_r977853794
##
tests/python/relay/test_op_level2.py:
##
@@ -2137,5 +2137,80 @@ def get_subgraph(dtype):
np.testing.assert_allclose(out, ref, rtol=1e-5, atol=1e-5)
Lunderberg commented on PR #12691:
URL: https://github.com/apache/tvm/pull/12691#issuecomment-1255253699
Rebased onto main, as https://github.com/apache/tvm/pull/12692 also needed
and implemented the same type-checking structs, which makes this PR much
smaller.
--
This is an automated
tkonolige commented on issue #12858:
URL: https://github.com/apache/tvm/issues/12858#issuecomment-1255234586
I'm fine with adding a flag to disable the warmup iteration, but warmup
iterations should be on by default. On CUDA, the first run on the model is very
slow due to lazy
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 39f71ae288 [frontend][pytorch] Add a new test case for torch
aten::fill_ operator implementation (#12857)
juda commented on PR #12318:
URL: https://github.com/apache/tvm/pull/12318#issuecomment-1255006044
Hi @masahi , I polished the tutorial according to your feedback, could you
please read it one more time?
--
This is an automated message from the Apache Git Service.
To respond to the
elvin-n commented on code in PR #12833:
URL: https://github.com/apache/tvm/pull/12833#discussion_r977607079
##
tests/python/relay/opencl_texture/test_conv2d_nchw_texture.py:
##
@@ -577,7 +577,6 @@ def test_residual_block(target, dtype):
"global.texture-weight",
chengven027-intellif opened a new pull request, #12872:
URL: https://github.com/apache/tvm/pull/12872
Hi, tvm:
Sorry,there is another operator `aten::zero_` just discovered that needs
to be reviewed.
--
This is an automated message from the Apache Git Service.
To respond to the
lhutton1 commented on issue #12858:
URL: https://github.com/apache/tvm/issues/12858#issuecomment-1254866832
Hi @f2013519, I submitted #12713 which makes sure the module is only run
once when the user is _not_ worried about benchmarking so wanted to share some
of my findings here. I also
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 39f71ae288 [frontend][pytorch] Add a new test case for torch
aten::fill_ operator implementation (#12857)
add
masahi merged PR #12545:
URL: https://github.com/apache/tvm/pull/12545
--
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:
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 7aef584c0f [Hybrid] Fix sys version check (#12837)
add 39f71ae288 [frontend][pytorch] Add a new test
elvin-n commented on PR #12545:
URL: https://github.com/apache/tvm/pull/12545#issuecomment-1254824500
can we merge this PR into main?
--
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
masahi closed issue #12870: multi-backends solution
URL: https://github.com/apache/tvm/issues/12870
--
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:
masahi commented on issue #12870:
URL: https://github.com/apache/tvm/issues/12870#issuecomment-1254740865
Please post such question to the discuss forum.
--
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
masahi closed issue #12844: [Bug][Frontend][PyTorch] Bug in the aten::fill_()
PyTorch operator implementation?
URL: https://github.com/apache/tvm/issues/12844
--
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
chayliu1991 opened a new issue, #12870:
URL: https://github.com/apache/tvm/issues/12870
There are multi-backends in my platform,include arm cpu+mali gpu+different
npus.
Can I use tvm to dynamic scheduling different backend?
--
This is an automated message from the Apache Git Service.
sfvaroglu commented on PR #12869:
URL: https://github.com/apache/tvm/pull/12869#issuecomment-1254710003
The roi_align test passes locally when using `onnx==1.12.0` and
`onnxruntime==1.12.1`, but it looks like CI is failing with
`onnxruntime.capi.onnxruntime_pybind11_state.InvalidGraph:
f2013519 commented on issue #12858:
URL: https://github.com/apache/tvm/issues/12858#issuecomment-1254694090
Also with --number=0, we get the below warning:
/usr/local/lib/python3.7/dist-packages/numpy/core/_methods.py:202:
RuntimeWarning: invalid value encountered in subtract x =
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 7aef584c0f [Hybrid] Fix sys version check (#12837)
add 39f71ae288 [frontend][pytorch] Add a new test case for torch
masahi merged PR #12857:
URL: https://github.com/apache/tvm/pull/12857
--
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:
sfvaroglu commented on PR #12869:
URL: https://github.com/apache/tvm/pull/12869#issuecomment-1254585495
@leandron @driazati
--
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.
sfvaroglu opened a new pull request, #12869:
URL: https://github.com/apache/tvm/pull/12869
RoiAlign-16 introduces `coordinate_transformation_mode`, which should be set
to 'output_half_pixel' to omit the pixel shift for the input (for a
backward-compatible behavior). This PR should fix the
Aympab opened a new pull request, #12868:
URL: https://github.com/apache/tvm/pull/12868
Following [this discussion](
https://discuss.tvm.apache.org/t/error-using-vm-to-profile-downcast-from-runtime-adt-to-runtime-ndarray-failed/13557/2)
from apache TVM discuss forums
The code of the
Suhong-Lee closed pull request #12867: Lwc
URL: https://github.com/apache/tvm/pull/12867
--
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:
75 matches
Mail list logo