[GitHub] [incubator-tvm] hello-hzb commented on issue #2954: [RFC][AUTOTVM] Auto-Schedule from Compute Declaration

hello-hzb commented on issue #2954: [RFC][AUTOTVM] Auto-Schedule from Compute Declaration URL: https://github.com/apache/incubator-tvm/issues/2954#issuecomment-552306982 @merrymercy Hi Zheng， My team has paied attension to your auto-scheduler work for a few days. No update for a

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule.

FrozenGene commented on a change in pull request #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule. URL: https://github.com/apache/incubator-tvm/pull/4277#discussion_r344568267 ## File path: topi/python/topi/arm_cpu/conv2d_spatial_pack.py ## @@ -93,24

[GitHub] [incubator-tvm] yzhliu commented on issue #4083: [tutorial] Relay pass infra tutorial

yzhliu commented on issue #4083: [tutorial] Relay pass infra tutorial URL: https://github.com/apache/incubator-tvm/pull/4083#issuecomment-552304594 Thanks @zhiics @billzorn @MarisaKirisame @yongwww This is an automated

[GitHub] [incubator-tvm] yzhliu merged pull request #4083: [tutorial] Relay pass infra tutorial

yzhliu merged pull request #4083: [tutorial] Relay pass infra tutorial URL: https://github.com/apache/incubator-tvm/pull/4083 This is an automated message from the Apache Git Service. To respond to the message, please log on

[incubator-tvm] branch master updated (1d24366 -> cff62bd)

This is an automated email from the ASF dual-hosted git repository. liuyizhi pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from 1d24366 [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. (#4249) add cff62bd

[GitHub] [incubator-tvm] jwfromm commented on a change in pull request #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

jwfromm commented on a change in pull request #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#discussion_r344565585 ## File path: tests/python/frontend/onnx/test_forward.py ## @@ -77,22

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule.

FrozenGene commented on a change in pull request #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule. URL: https://github.com/apache/incubator-tvm/pull/4277#discussion_r344559034 ## File path: topi/python/topi/arm_cpu/conv2d_spatial_pack.py ## @@ -93,24

[GitHub] [incubator-tvm] cchung100m commented on a change in pull request #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

cchung100m commented on a change in pull request #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#discussion_r344558194 ## File path: tests/python/frontend/onnx/test_forward.py ## @@ -85,6

[GitHub] [incubator-tvm] yongwww opened a new pull request #4301: [TF][TEST] add test_forward_reduce_any back

yongwww opened a new pull request #4301: [TF][TEST] add test_forward_reduce_any back URL: https://github.com/apache/incubator-tvm/pull/4301 the test case was removed in #4181 for some reason @tqchen @soiferj @zhiics

[GitHub] [incubator-tvm] jwfromm commented on a change in pull request #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

jwfromm commented on a change in pull request #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#discussion_r344549188 ## File path: tests/python/frontend/onnx/test_forward.py ## @@ -85,6

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

tqchen commented on a change in pull request #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#discussion_r344549075 ## File path: include/tvm/runtime/module.h ## @@ -71,7 +70,11 @@ class

[GitHub] [incubator-tvm] tqchen commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

tqchen commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#issuecomment-552281875 after thinking a bit more, i agree that ObjectInternal could indeed be a better name, I just pushed a change to

[GitHub] [incubator-tvm] tqchen commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

tqchen commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#issuecomment-552280462 I thought a bit about Object::Internal, but eventually chose the name FFIUtil because it also has a name for

[GitHub] [incubator-tvm] jwfromm edited a comment on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

jwfromm edited a comment on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#issuecomment-552280147 I've added onnxruntime to the docker image in PR #4299. Once it's merged and the CI image is rebuilt

[GitHub] [incubator-tvm] jwfromm commented on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

jwfromm commented on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#issuecomment-552280147 I've added onnxruntime to the docker image in PR #4299. Once it's merged the tests in this PR should be able

[GitHub] [incubator-tvm] jwfromm opened a new pull request #4299: [CI] Add onnxruntime to CI Image

jwfromm opened a new pull request #4299: [CI] Add onnxruntime to CI Image URL: https://github.com/apache/incubator-tvm/pull/4299 As discussed in PR #4271, using the caffe runtime for our onnx testing is not a good long term solution due to lack of op support. Instead, we should be using

[GitHub] [incubator-tvm] junrushao1994 commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

junrushao1994 commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#issuecomment-552277277 @tqchen The use case of the class `FFIUtil` makes sense to me now, but I think it would be great to have

[GitHub] [incubator-tvm] yzhliu commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation.

yzhliu commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. URL: https://github.com/apache/incubator-tvm/pull/4249#issuecomment-552275667 Thanks @anijain2305 @FrozenGene @jackwish @tqchen

[GitHub] [incubator-tvm] yzhliu merged pull request #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation.

yzhliu merged pull request #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. URL: https://github.com/apache/incubator-tvm/pull/4249 This is an automated message from the Apache Git Service. To

[incubator-tvm] branch master updated: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. (#4249)

This is an automated email from the ASF dual-hosted git repository. liuyizhi 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 1d24366 [TOPI][AlterOpLayout][ARM]

[GitHub] [incubator-tvm] yzhliu edited a comment on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate

yzhliu edited a comment on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-550417287 # Build and Test * Increate the robuteness of CI test ([#2841](https://github.com/dmlc/tvm/pull/2841),

[GitHub] [incubator-tvm] yzhliu removed a comment on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate

yzhliu removed a comment on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-552266647 # Build and Test * Increate the robuteness of CI test ([#2841](https://github.com/dmlc/tvm/pull/2841),

[GitHub] [incubator-tvm] yzhliu commented on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate

yzhliu commented on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-552266647 # Build and Test * Increate the robuteness of CI test ([#2841](https://github.com/dmlc/tvm/pull/2841),

[GitHub] [incubator-tvm] jackwish commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation.

jackwish commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. URL: https://github.com/apache/incubator-tvm/pull/4249#issuecomment-552263685 That sounds good, thank you :) @anijain2305

[GitHub] [incubator-tvm] tqchen commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

tqchen commented on issue #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#issuecomment-552255958 The main reason for the ffi_util.h is to keep these functionalities internal, as we do not want to expose these

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

junrushao1994 commented on a change in pull request #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#discussion_r344525003 ## File path: src/runtime/micro/tcl_socket.h ## @@ -18,7 +18,6 @@

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module

junrushao1994 commented on a change in pull request #4289: [RUNTIME][REFACTOR] Use object protocol to support runtime::Module URL: https://github.com/apache/incubator-tvm/pull/4289#discussion_r344524435 ## File path: include/tvm/runtime/module.h ## @@ -71,7 +70,11 @@

[incubator-tvm] branch master updated (2f65a87 -> d2fc025)

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 2f65a87 [TFLite] Support PRelu (#4298) add d2fc025 [RUTNIME] Support C++ RPC (#4281) No new revisions

[GitHub] [incubator-tvm] tqchen commented on issue #4281: [RUTNIME] Support C++ RPC

tqchen commented on issue #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#issuecomment-552244083 Thanks @FrozenGene @weberlo ! This is an automated message from the Apache Git Service.

[GitHub] [incubator-tvm] tqchen merged pull request #4281: [RUTNIME] Support C++ RPC

tqchen merged pull request #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[GitHub] [incubator-tvm] junrushao1994 commented on issue #4275: imp module is deprecated

junrushao1994 commented on issue #4275: imp module is deprecated URL: https://github.com/apache/incubator-tvm/pull/4275#issuecomment-552227157 Seems like your PR failed on an irrelevant place, so feel free to re-trigger it. BTW, I saw there is usage of `import imp` as well in

[incubator-tvm] branch master updated (fc28f7a -> 2f65a87)

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 fc28f7a [Test][TF][Relay] Fix argument preparation for vm test mode (#4296) add 2f65a87 [TFLite]

[GitHub] [incubator-tvm] tqchen commented on issue #4298: [TFLite] Support PRelu

tqchen commented on issue #4298: [TFLite] Support PRelu URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-552226928 Thanks @FrozenGene This is an automated message from the Apache Git Service. To respond to

[GitHub] [incubator-tvm] tqchen merged pull request #4298: [TFLite] Support PRelu

tqchen merged pull request #4298: [TFLite] Support PRelu URL: https://github.com/apache/incubator-tvm/pull/4298 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and

[GitHub] [incubator-tvm] anijain2305 commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation.

anijain2305 commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. URL: https://github.com/apache/incubator-tvm/pull/4249#issuecomment-55356 > Thanks for your detailed explanation @anijain2305 ! I missed you change of removing legalization.

[incubator-tvm] branch master updated (801cf0e -> fc28f7a)

This is an automated email from the ASF dual-hosted git repository. zhic pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from 801cf0e [Codegen][cuda-fp16] fallback to fp32 simulation when cuda arch < sm53 (#4268) add fc28f7a

[GitHub] [incubator-tvm] zhiics merged pull request #4296: [Test][TF][Relay] Fix argument preparation for vm test mode

zhiics merged pull request #4296: [Test][TF][Relay] Fix argument preparation for vm test mode URL: https://github.com/apache/incubator-tvm/pull/4296 This is an automated message from the Apache Git Service. To respond to

[GitHub] [incubator-tvm] tqchen commented on issue #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule.

tqchen commented on issue #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule. URL: https://github.com/apache/incubator-tvm/pull/4277#issuecomment-552219578 @ajtulloch it would be great if you can manage this PR, you should have the permission to approve and merge

[GitHub] [incubator-tvm] tqchen commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation.

tqchen commented on issue #4249: [TOPI][AlterOpLayout][ARM] Enabling NHWC to NCHW layout transformation. URL: https://github.com/apache/incubator-tvm/pull/4249#issuecomment-552219192 I think we want an explicit LayoutConvert pass besides the automatic one to give developer more

[GitHub] [incubator-tvm] FrozenGene commented on issue #4281: [RUTNIME] Support C++ RPC

FrozenGene commented on issue #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#issuecomment-552219131 @tqchen Have resolved LOG(ERROR) -> LOG(FATAL) and gnulib issue. This is an

[GitHub] [incubator-tvm] FrozenGene edited a comment on issue #4281: [RUTNIME] Support C++ RPC

FrozenGene edited a comment on issue #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#issuecomment-552219131 @tqchen Have resolved LOG(ERROR) -> LOG(FATAL) and gnulib issue. Please review again

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC

FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#discussion_r344506405 ## File path: src/common/util.h ## @@ -61,6 +65,41 @@ inline int TVMPClose(FILE* stream) { return

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC

FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#discussion_r344506400 ## File path: apps/cpp_rpc/rpc_env.cc ## @@ -164,8 +163,11 @@ void LinuxShared(const std::string output,

[GitHub] [incubator-tvm] tqchen commented on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

tqchen commented on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#issuecomment-552218937 The procedure will be add onnxruntime dep to the docker/install/install_ubuntu_onnx.sh, then we will rebuild

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #4281: [RUTNIME] Support C++ RPC

tqchen commented on a change in pull request #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#discussion_r344506148 ## File path: src/common/util.h ## @@ -61,6 +65,41 @@ inline int TVMPClose(FILE* stream) { return pclose(stream);

[GitHub] [incubator-tvm] FrozenGene opened a new pull request #4298: [TFLite] Support PRelu

FrozenGene opened a new pull request #4298: [TFLite] Support PRelu URL: https://github.com/apache/incubator-tvm/pull/4298 As discuss forum topic: https://discuss.tvm.ai/t/missing-tflite-operators/3150/8, we want to have PRelu support in TFLite @apivovarov

[GitHub] [incubator-tvm] icemelon9 commented on a change in pull request #4158: [PERF] Parallelize reduction for CPU

icemelon9 commented on a change in pull request #4158: [PERF] Parallelize reduction for CPU URL: https://github.com/apache/incubator-tvm/pull/4158#discussion_r344504236 ## File path: topi/python/topi/x86/reduction.py ## @@ -0,0 +1,81 @@ +# Licensed to the Apache Software

[GitHub] [incubator-tvm] icemelon9 commented on a change in pull request #4158: [PERF] Parallelize reduction for CPU

icemelon9 commented on a change in pull request #4158: [PERF] Parallelize reduction for CPU URL: https://github.com/apache/incubator-tvm/pull/4158#discussion_r344504230 ## File path: topi/python/topi/x86/reduction.py ## @@ -0,0 +1,81 @@ +# Licensed to the Apache Software

[GitHub] [incubator-tvm] cchung100m commented on issue #4283: Add support for quant. mul operator in tflite frontend

cchung100m commented on issue #4283: Add support for quant. mul operator in tflite frontend URL: https://github.com/apache/incubator-tvm/pull/4283#issuecomment-552193779 Hi @inadob Should we add the test case for `quant.mul `operator in tflite frontend?

[GitHub] [incubator-tvm] FrozenGene opened a new pull request #4297: [OP] Support Faster-RCNN Proposal OP on CPU.

FrozenGene opened a new pull request #4297: [OP] Support Faster-RCNN Proposal OP on CPU. URL: https://github.com/apache/incubator-tvm/pull/4297 Currently, our TVM only implement proposal op on CUDA. However, we should support Proposal op on CPU too. @vinx13 @tqchen please help to

[GitHub] [incubator-tvm] FrozenGene commented on issue #4281: [RUTNIME] Support C++ RPC

FrozenGene commented on issue #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#issuecomment-552174191 @tqchen @weberlo I should resolve all the comments. Please review again. This is an

[GitHub] [incubator-tvm] jackwish edited a comment on issue #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule.

jackwish edited a comment on issue #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule. URL: https://github.com/apache/incubator-tvm/pull/4277#issuecomment-552174176 Thanks for your kind explanation @anijain2305 . That is exactly my consideration as the legalization approach

[GitHub] [incubator-tvm] jackwish commented on issue #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule.

jackwish commented on issue #4277: [ARM][Topi] Improving Int8 Perf in Spatial Conv2D schedule. URL: https://github.com/apache/incubator-tvm/pull/4277#issuecomment-552174176 Thanks for your kind explanation @anijain2305 . That is exactly my consideration as the legalization approach may

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC

FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#discussion_r344478767 ## File path: apps/cpp_rpc/rpc_env.cc ## @@ -0,0 +1,247 @@ +/* + * Licensed to the Apache Software Foundation

[GitHub] [incubator-tvm] FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC

FrozenGene commented on a change in pull request #4281: [RUTNIME] Support C++ RPC URL: https://github.com/apache/incubator-tvm/pull/4281#discussion_r344478615 ## File path: src/runtime/rpc/rpc_session.h ## @@ -36,8 +36,29 @@ namespace tvm { namespace runtime { +//

[GitHub] [incubator-tvm] cchung100m commented on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth

cchung100m commented on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#issuecomment-552173389 Hi @jwfromm Thanks for the prompt reply. I agree that we can use the `onnxruntime` to test