[GitHub] [incubator-tvm] mkstarr commented on pull request #6580: Faster sparse_dense on GPUs

mkstarr commented on pull request #6580: URL: https://github.com/apache/incubator-tvm/pull/6580#issuecomment-701185298 Have you considered your syntax errors? This is an automated message from the Apache Git Service. To respo

[GitHub] [incubator-tvm] ANSHUMAN87 commented on pull request #6580: Faster sparse_dense on GPUs

ANSHUMAN87 commented on pull request #6580: URL: https://github.com/apache/incubator-tvm/pull/6580#issuecomment-701169720 @tkonolige : I understand your concern clearly. However it was just a thought. Even if run to run or machine to machine difference, the relative reference would be same

[GitHub] [incubator-tvm] masahi commented on issue #6535: [Torch] Support aten::tensor, aten::empty and aten::numel

masahi commented on issue #6535: URL: https://github.com/apache/incubator-tvm/issues/6535#issuecomment-701127623 ah sorry, the link is fixed This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [incubator-tvm] masahi edited a comment on issue #6535: [Torch] Support aten::tensor, aten::empty and aten::numel

masahi edited a comment on issue #6535: URL: https://github.com/apache/incubator-tvm/issues/6535#issuecomment-701110902 We'll track 1.6 support at https://github.com/apache/incubator-tvm/issues/6594 This is an automated mess

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

FrozenGene commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r497209168 ## File path: tests/python/driver/tvmc/conftest.py ## @@ -110,10 +112,33 @@ def onnx_resnet50(): @pytest.fixture(scope="session") def tflite_

[GitHub] [incubator-tvm] zhiqwang commented on issue #6535: [Torch] Support aten::tensor, aten::empty and aten::numel

zhiqwang commented on issue #6535: URL: https://github.com/apache/incubator-tvm/issues/6535#issuecomment-701121269 Hi, That's great news! Relink it to #6594 ? This is an automated message from the Apache Git Service. To respo

[GitHub] [incubator-tvm] masahi commented on issue #6535: [Torch] Support aten::tensor, aten::empty and aten::numel

masahi commented on issue #6535: URL: https://github.com/apache/incubator-tvm/issues/6535#issuecomment-701110902 We'll track 1.6 support at https://github.com/apache/incubator-tvm/issues/6535 This is an automated message fro

[GitHub] [incubator-tvm] masahi closed issue #6535: [Torch] Support aten::tensor, aten::empty and aten::numel

masahi closed issue #6535: URL: https://github.com/apache/incubator-tvm/issues/6535 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

[GitHub] [incubator-tvm] masahi opened a new issue #6594: [Torch, CI] Support PyTorch 1.6

masahi opened a new issue #6594: URL: https://github.com/apache/incubator-tvm/issues/6594 Let's summarize steps needed to update our PyTorch support for the latest version, 1.6 (the CI is at 1.4). - [ ] Quantization support: It is completely broken due to a bug introduced in 1.6 (s

[GitHub] [incubator-tvm] imalsogreg commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

imalsogreg commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497193474 ## File path: rust/tvm/examples/resnet/build.rs ## @@ -17,16 +17,20 @@ * under the License. */ -use anyhow::{Context, Result}; -use std::{p

[GitHub] [incubator-tvm] imalsogreg commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

imalsogreg commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497191902 ## File path: rust/tvm/Cargo.toml ## @@ -40,6 +40,10 @@ tvm-macros = { version = "*", path = "../tvm-macros/" } paste = "0.1" mashup = "0.1" o

[GitHub] [incubator-tvm] imalsogreg commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

imalsogreg commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497180923 ## File path: rust/tvm-rt/src/ndarray.rs ## @@ -184,28 +246,19 @@ impl NDArray { /// let ctx = Context::cpu(0); /// let mut ndarray = N

[GitHub] [incubator-tvm] imalsogreg commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

imalsogreg commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497171514 ## File path: rust/tvm-rt/src/ndarray.rs ## @@ -374,28 +430,27 @@ mod tests { #[test] fn basics() { -let shape = &mut [1, 2,

[GitHub] [incubator-tvm] jroesch commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

jroesch commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497170872 ## File path: rust/tvm-rt/src/ndarray.rs ## @@ -184,28 +246,19 @@ impl NDArray { /// let ctx = Context::cpu(0); /// let mut ndarray = NDAr

[GitHub] [incubator-tvm] ZihengJiang commented on issue #6486: TVM v0.7 Release Note Candidate

ZihengJiang commented on issue #6486: URL: https://github.com/apache/incubator-tvm/issues/6486#issuecomment-701059777 I will add an `API Changes` section before `Deprecation`. This is an automated message from the Apache Git

[GitHub] [incubator-tvm] ZihengJiang edited a comment on issue #6486: TVM v0.7 Release Note Candidate

ZihengJiang edited a comment on issue #6486: URL: https://github.com/apache/incubator-tvm/issues/6486#issuecomment-701059777 I will add an `API Changes` section before `Deprecation`. @roastduck @tqchen This is an automated m

[GitHub] [incubator-tvm] imalsogreg commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

imalsogreg commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497168051 ## File path: rust/tvm-rt/src/ndarray.rs ## @@ -184,28 +246,19 @@ impl NDArray { /// let ctx = Context::cpu(0); /// let mut ndarray = N

[GitHub] [incubator-tvm] trevor-m commented on pull request #6395: [BYOC][TensorRT] TensorRT BYOC integration

trevor-m commented on pull request #6395: URL: https://github.com/apache/incubator-tvm/pull/6395#issuecomment-701059299 > @trevor-m i agree this seems unrelated. can you try retriggering the CI just to double-check it's reproducible? there was an issue with flaky bundle_deploy tests a whil

[GitHub] [incubator-tvm] imalsogreg commented on a change in pull request #6563: [Rust] Improve NDArray, GraphRt, and Relay bindings

imalsogreg commented on a change in pull request #6563: URL: https://github.com/apache/incubator-tvm/pull/6563#discussion_r497162386 ## File path: rust/tvm-rt/src/ndarray.rs ## @@ -47,73 +47,146 @@ //! [`copy_from_buffer`]:struct.NDArray.html#method.copy_from_buffer //! [`cop

[GitHub] [incubator-tvm] areusch commented on pull request #6395: [BYOC][TensorRT] TensorRT BYOC integration

areusch commented on pull request #6395: URL: https://github.com/apache/incubator-tvm/pull/6395#issuecomment-701055475 @trevor-m i agree this seems unrelated. can you try retriggering the CI just to double-check it's reproducible? there was an issue with flaky bundle_deploy tests a while b

[GitHub] [incubator-tvm] areusch commented on issue #5298: [CRT][FLAKY] Flaky failures on 32bit platforms

areusch commented on issue #5298: URL: https://github.com/apache/incubator-tvm/issues/5298#issuecomment-701055150 Possibly another instance spotted in #6395, but can't view the old jenkins logs to check: https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-6395/26/pipeline

[GitHub] [incubator-tvm] trevor-m edited a comment on pull request #6395: [BYOC][TensorRT] TensorRT BYOC integration

trevor-m edited a comment on pull request #6395: URL: https://github.com/apache/incubator-tvm/pull/6395#issuecomment-701036415 It looks like `/tests/scripts/task_cpp_unittest.sh` is failing at "Test MISRA-C runtime" in the CI: ``` python3 build_model.py -o build --test INFO:compile

[GitHub] [incubator-tvm] trevor-m commented on pull request #6395: [BYOC][TensorRT] TensorRT BYOC integration

trevor-m commented on pull request #6395: URL: https://github.com/apache/incubator-tvm/pull/6395#issuecomment-701036415 It looks like `/tests/scripts/task_cpp_unittest.sh` is failing at "Test MISRA-C runtime": ``` python3 build_model.py -o build --test INFO:compile_engine:Using inj

[GitHub] [incubator-tvm] tkonolige commented on pull request #6580: Faster sparse_dense on GPUs

tkonolige commented on pull request #6580: URL: https://github.com/apache/incubator-tvm/pull/6580#issuecomment-701024445 @ANSHUMAN87 Right now TVM does not do any testing for performance regressions. The hard part in setting up performance testing is that is varies from run to run and mach

[GitHub] [incubator-tvm] jwfromm opened a new pull request #6593: [Topi][Cuda] Tiny bug fix for non-fp32 datatypes in conv2d_transpose.

jwfromm opened a new pull request #6593: URL: https://github.com/apache/incubator-tvm/pull/6593 The pad operator in the Cuda definition of conv2d_transpose assumed the input type was `float32`. This oneliner PR fixes that assumption. ---

[GitHub] [incubator-tvm] tqchen commented on issue #6592: [Docs] relay.nn.bias_add with axis=-1?

tqchen commented on issue #6592: URL: https://github.com/apache/incubator-tvm/issues/6592#issuecomment-701001050 I agree it might be helpful to update the document, Thanks @gussmith23 for bringing this up This is an automate

[GitHub] [incubator-tvm] trevor-m commented on a change in pull request #6395: [BYOC][TensorRT] TensorRT BYOC integration

trevor-m commented on a change in pull request #6395: URL: https://github.com/apache/incubator-tvm/pull/6395#discussion_r497068599 ## File path: src/runtime/contrib/tensorrt/tensorrt_runtime.cc ## @@ -0,0 +1,311 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under

[GitHub] [incubator-tvm] gussmith23 commented on issue #6592: [Docs] relay.nn.bias_add with axis=-1?

gussmith23 commented on issue #6592: URL: https://github.com/apache/incubator-tvm/issues/6592#issuecomment-700998557 It is somewhat implicit what negative `axis` values mean; they're used in other places in Relay. I bet a lot of people know what it means right off the bat. --

[GitHub] [incubator-tvm] tqchen closed issue #6592: [Docs] relay.nn.bias_add with axis=-1?

tqchen closed issue #6592: URL: https://github.com/apache/incubator-tvm/issues/6592 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

[GitHub] [incubator-tvm] tqchen commented on issue #6592: [Docs] relay.nn.bias_add with axis=-1?

tqchen commented on issue #6592: URL: https://github.com/apache/incubator-tvm/issues/6592#issuecomment-700975858 I believe it is like most numpy conventions. Negative axis means starting from the highest dimension. Let us also use https://discuss.tvm.apache.org/ for this kind of gene

[incubator-tvm] branch master updated (c0a6bc3 -> 0535fd1)

This is an automated email from the ASF dual-hosted git repository. tqchen pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from c0a6bc3 [tvmc] Introduce 'tune' subcommand (part 3/4) (#6537) add 0535fd1 [RUNTIME] NDArray CopyFrom/To

[GitHub] [incubator-tvm] tqchen commented on pull request #6586: [RUNTIME] NDArray CopyFrom/To Bytes always synchronize

tqchen commented on pull request #6586: URL: https://github.com/apache/incubator-tvm/pull/6586#issuecomment-700974264 Thanks @tmoreau89 @zhiics @ANSHUMAN87 @ZihengJiang This is an automated message from the Apache Git Servic

[GitHub] [incubator-tvm] tqchen merged pull request #6586: [RUNTIME] NDArray CopyFrom/To Bytes always synchronize

tqchen merged pull request #6586: URL: https://github.com/apache/incubator-tvm/pull/6586 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] [incubator-tvm] tqchen commented on a change in pull request #6586: [RUNTIME] NDArray CopyFrom/To Bytes always synchronize

tqchen commented on a change in pull request #6586: URL: https://github.com/apache/incubator-tvm/pull/6586#discussion_r497034108 ## File path: src/runtime/ndarray.cc ## @@ -70,9 +70,12 @@ void ArrayCopyFromBytes(DLTensor* handle, const void* data, size_t nbytes) { cpu_ctx.d

[GitHub] [incubator-tvm] gussmith23 commented on issue #6592: [Docs] relay.nn.bias_add with axis=-1?

gussmith23 commented on issue #6592: URL: https://github.com/apache/incubator-tvm/issues/6592#issuecomment-700953127 cc @hogepodge This is an automated message from the Apache Git Service. To respond to the message, please l

[GitHub] [incubator-tvm] gussmith23 opened a new issue #6592: [Docs] relay.nn.bias_add with axis=-1?

gussmith23 opened a new issue #6592: URL: https://github.com/apache/incubator-tvm/issues/6592 The documentation for `bias_add` does not seem to indicate what `axis=-1` means, though it appears in Relay's version of ResNet. T

[GitHub] [incubator-tvm] ANSHUMAN87 commented on pull request #6580: Faster sparse_dense on GPUs

ANSHUMAN87 commented on pull request #6580: URL: https://github.com/apache/incubator-tvm/pull/6580#issuecomment-700929567 > I've written a faster sparse_dense for GPUs using tir. This sparse_dense requires a padded matrix, so I've added a new op sparse_dense_padded. AlterOpLayout should tr

[GitHub] [incubator-tvm] ANSHUMAN87 commented on a change in pull request #6586: [RUNTIME] NDArray CopyFrom/To Bytes always synchronize

ANSHUMAN87 commented on a change in pull request #6586: URL: https://github.com/apache/incubator-tvm/pull/6586#discussion_r496975400 ## File path: src/runtime/ndarray.cc ## @@ -70,9 +70,12 @@ void ArrayCopyFromBytes(DLTensor* handle, const void* data, size_t nbytes) { cpu_c

[GitHub] [incubator-tvm] tkonolige commented on pull request #6591: [FIX] Use cmake to dispatch to the build tool

tkonolige commented on pull request #6591: URL: https://github.com/apache/incubator-tvm/pull/6591#issuecomment-700907706 It doesn't look like it is easy to pass the number of jobs from the makefile to cmake. This is an autom

[GitHub] [incubator-tvm] tkonolige closed pull request #6591: [FIX] Use cmake to dispatch to the build tool

tkonolige closed pull request #6591: URL: https://github.com/apache/incubator-tvm/pull/6591 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] [incubator-tvm] areusch commented on a change in pull request #6485: Add ci_qemu docker image

areusch commented on a change in pull request #6485: URL: https://github.com/apache/incubator-tvm/pull/6485#discussion_r496954365 ## File path: docker/install/ubuntu_install_qemu.sh ## @@ -20,4 +20,473 @@ set -e set -u set -o pipefail -apt-get -y install qemu-system +sudo s

[GitHub] [incubator-tvm] tqchen commented on pull request #6591: [FIX] Use cmake to dispatch to the build tool

tqchen commented on pull request #6591: URL: https://github.com/apache/incubator-tvm/pull/6591#issuecomment-700900342 This might will create some regression behavior since it is common for users to type say `make -j9`, would be great if we can pass argument along -

[GitHub] [incubator-tvm] tkonolige opened a new pull request #6591: [FIX] Use cmake to dispatch to the build tool

tkonolige opened a new pull request #6591: URL: https://github.com/apache/incubator-tvm/pull/6591 Fix the Makefile to use cmake to run the build tool. This lets users use alternative build tools like ninja. @jroesch T

[GitHub] [incubator-tvm] comaniac commented on a change in pull request #6584: [Ansor] Support multiple output ops and fix Python API printing

comaniac commented on a change in pull request #6584: URL: https://github.com/apache/incubator-tvm/pull/6584#discussion_r496369109 ## File path: src/auto_scheduler/compute_dag.cc ## @@ -969,6 +969,9 @@ void ComputeDAG::RewriteLayout(const Array& transform_steps) { }

[GitHub] [incubator-tvm] tqchen closed issue #6590: How to free memory manually after building a operator for long running building case?

tqchen closed issue #6590: URL: https://github.com/apache/incubator-tvm/issues/6590 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

[GitHub] [incubator-tvm] tqchen commented on issue #6590: How to free memory manually after building a operator for long running building case?

tqchen commented on issue #6590: URL: https://github.com/apache/incubator-tvm/issues/6590#issuecomment-700875357 Thanks for reporting the problem, please open a new thread on https://discuss.tvm.apache.org/. BTW, checking a modified version of the script on master seems to show a stable me

[GitHub] [incubator-tvm] comaniac commented on pull request #6537: [tvmc] Introduce 'tune' subcommand (part 3/4)

comaniac commented on pull request #6537: URL: https://github.com/apache/incubator-tvm/pull/6537#issuecomment-700837093 Thanks @leandron @FrozenGene This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-tvm] comaniac merged pull request #6537: [tvmc] Introduce 'tune' subcommand (part 3/4)

comaniac merged pull request #6537: URL: https://github.com/apache/incubator-tvm/pull/6537 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

[incubator-tvm] branch master updated (27abfad -> c0a6bc3)

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from 27abfad disable stacked bidir test (#6585) add c0a6bc3 [tvmc] Introduce 'tune' subcommand (part 3/4)

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

mbaret commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496883799 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contribu

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

mbaret commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496881232 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contribu

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

mbaret commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496880495 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contribu

[GitHub] [incubator-tvm] tkonolige commented on a change in pull request #6274: [Diagnostics][Relay][InferType] Refactor InferType to work on whole module, and use new diagnostics.

tkonolige commented on a change in pull request #6274: URL: https://github.com/apache/incubator-tvm/pull/6274#discussion_r496878564 ## File path: python/tvm/relay/prelude.py ## @@ -259,33 +284,40 @@ def define_tensor_array_read(self): Tensor[self.shape, self.dtype]

[GitHub] [incubator-tvm] tkonolige commented on a change in pull request #6274: [Diagnostics][Relay][InferType] Refactor InferType to work on whole module, and use new diagnostics.

tkonolige commented on a change in pull request #6274: URL: https://github.com/apache/incubator-tvm/pull/6274#discussion_r496874950 ## File path: Makefile ## @@ -135,6 +135,11 @@ jvminstall: mvn install -P$(JVM_PKG_PROFILE) -Dcxx="$(CXX)" \

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496873622 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] csullivan commented on pull request #6579: [ROCM] Support rocblas_sgemm_strided_batched

csullivan commented on pull request #6579: URL: https://github.com/apache/incubator-tvm/pull/6579#issuecomment-700814058 Thank you @masahi ! This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [incubator-tvm] siju-samuel commented on a change in pull request #6589: [Frontend][Tensorflow] Fix TF 1.15 conv2d_transpose parsing

siju-samuel commented on a change in pull request #6589: URL: https://github.com/apache/incubator-tvm/pull/6589#discussion_r496856859 ## File path: python/tvm/relay/frontend/tensorflow.py ## @@ -302,7 +302,11 @@ def _impl(inputs, attr, params, mod): )

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496846703 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496830062 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496830062 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496827216 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

tqchen commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496808406 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contribu

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

tqchen commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496808406 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contribu

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #6485: Add ci_qemu docker image

tqchen commented on a change in pull request #6485: URL: https://github.com/apache/incubator-tvm/pull/6485#discussion_r496804408 ## File path: docker/install/ubuntu_install_qemu.sh ## @@ -20,4 +20,473 @@ set -e set -u set -o pipefail -apt-get -y install qemu-system +sudo se

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

FrozenGene commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496800421 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more cont

[GitHub] [incubator-tvm] tqchen commented on pull request #6586: [RUNTIME] NDArray CopyFrom/To Bytes always synchronize

tqchen commented on pull request #6586: URL: https://github.com/apache/incubator-tvm/pull/6586#issuecomment-700772795 Note that the async behavior can still be fully recovered by Allocating a CPU NDArray and copy from there. The current behavior only affects copyfromto bytes

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496796368 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #6445: Add dot product support for quantized convolution.

mbaret commented on a change in pull request #6445: URL: https://github.com/apache/incubator-tvm/pull/6445#discussion_r496745509 ## File path: python/tvm/topi/arm_cpu/tensor_intrin.py ## @@ -589,6 +587,236 @@ def _instr(index): ) +def select_word(vec, lane, dtype_vec):

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

FrozenGene commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496788964 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more cont

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496784763 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496781556 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496781713 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496773214 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496771378 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

FrozenGene commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496767403 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more cont

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496755746 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496755746 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496753128 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

FrozenGene commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496740342 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more cont

[GitHub] [incubator-tvm] jamesqiao opened a new issue #6590: How to free memory manually after building a operator for long running building case?

jamesqiao opened a new issue #6590: URL: https://github.com/apache/incubator-tvm/issues/6590 Hello all, I have a sample code from tvm offical site below. I want to build thousands of operators in the same process. But i found there is memory leak and the memory usage is increasin

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496732677 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496731437 ## File path: python/tvm/driver/tvmc/runner.py ## @@ -0,0 +1,450 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contri

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496728910 ## File path: python/tvm/driver/tvmc/common.py ## @@ -63,3 +72,32 @@ def convert_graph_layout(mod, desired_layout): raise TVMCExceptio

[GitHub] [incubator-tvm] leandron commented on a change in pull request #6578: [tvmc] Introduce 'run' subcommand (part 4/4)

leandron commented on a change in pull request #6578: URL: https://github.com/apache/incubator-tvm/pull/6578#discussion_r496727105 ## File path: python/tvm/driver/tvmc/common.py ## @@ -17,10 +17,19 @@ """ Common utility functions shared by TVMC modules. """ +import logging +

[GitHub] [incubator-tvm] leandron commented on pull request #6537: [tvmc] Introduce 'tune' subcommand (part 3/4)

leandron commented on pull request #6537: URL: https://github.com/apache/incubator-tvm/pull/6537#issuecomment-700701519 This PR is passing CI again. @comaniac @FrozenGene do you mind having another look? This is an aut

[GitHub] [incubator-tvm] lhutton1 commented on a change in pull request #6532: [BYOC][ACL] Support add operation

lhutton1 commented on a change in pull request #6532: URL: https://github.com/apache/incubator-tvm/pull/6532#discussion_r496634223 ## File path: tests/python/contrib/test_arm_compute_lib/test_add.py ## @@ -0,0 +1,135 @@ +# Licensed to the Apache Software Foundation (ASF) under

[GitHub] [incubator-tvm] jtuyls opened a new pull request #6589: [Frontend][Tensorflow] Fix TF 1.15 conv2d_transpose parsing

jtuyls opened a new pull request #6589: URL: https://github.com/apache/incubator-tvm/pull/6589 Fix for Tensorflow 1.15.0 conv2d_transpose parsing. Specifically, attr["_output_shapes"][0] can be None. Following snippet can be used to create a model for reproducing the bug (run with T

[GitHub] [incubator-tvm] masahi closed issue #6457: [TEST][FLAKY] tests/python/frontend/pytorch/test_lstm.py

masahi closed issue #6457: URL: https://github.com/apache/incubator-tvm/issues/6457 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

[incubator-tvm] branch master updated: Add rocblas_sgemm_strided_batched impl. (#6579)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git The following commit(s) were added to refs/heads/master by this push: new 22ad0dd Add rocblas_sgemm_strided_batche

[GitHub] [incubator-tvm] masahi commented on pull request #6579: [ROCM] Support rocblas_sgemm_strided_batched

masahi commented on pull request #6579: URL: https://github.com/apache/incubator-tvm/pull/6579#issuecomment-700600299 thanks @csullivan This is an automated message from the Apache Git Service. To respond to the message, ple

[GitHub] [incubator-tvm] masahi merged pull request #6585: [Test, Torch] Disable a flaky test

masahi merged pull request #6585: URL: https://github.com/apache/incubator-tvm/pull/6585 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] [incubator-tvm] masahi merged pull request #6579: [ROCM] Support rocblas_sgemm_strided_batched

masahi merged pull request #6579: URL: https://github.com/apache/incubator-tvm/pull/6579 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

[incubator-tvm] branch master updated: disable stacked bidir test (#6585)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a commit to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git The following commit(s) were added to refs/heads/master by this push: new 27abfad disable stacked bidir test (#658

[GitHub] [incubator-tvm] XIAO-XIA opened a new issue #6588: [Relay][BUG] nn.dense accuracy error

XIAO-XIA opened a new issue #6588: URL: https://github.com/apache/incubator-tvm/issues/6588 The output of relay op nn.dense seems not so accurate when working on some specific input shapes. `data_shape = (8, 512), weight_shape = (128, 512)` ### Reproduce ```python impor

[GitHub] [incubator-tvm] manupa-arm commented on a change in pull request #6560: [BYOC][ETHOSN] Fix tests for new module API

manupa-arm commented on a change in pull request #6560: URL: https://github.com/apache/incubator-tvm/pull/6560#discussion_r496564570 ## File path: tests/python/contrib/test_ethosn/test_networks.py ## @@ -104,17 +96,16 @@ def test_mobilenet_v1(): # codegen, which could come

[incubator-tvm] branch master updated (fde8d43 -> cab4ce1)

This is an automated email from the ASF dual-hosted git repository. ziheng pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from fde8d43 [tvmc] unify all logs on a single logger 'TVMC' (#6577) add cab4ce1 [COMMUNITY] vegaluisjose ->

[GitHub] [incubator-tvm] ZihengJiang merged pull request #6582: [COMMUNITY] vegaluisjose -> Committer

ZihengJiang merged pull request #6582: URL: https://github.com/apache/incubator-tvm/pull/6582 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