[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4275: imp module is deprecated

junrushao1994 commented on a change in pull request #4275: imp module is deprecated URL: https://github.com/apache/incubator-tvm/pull/4275#discussion_r346697415 ## File path: topi/python/topi/cpp/vision/__init__.py ## @@ -0,0 +1,7 @@ +"""FFI for vision TOPI ops and

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4275: imp module is deprecated

junrushao1994 commented on a change in pull request #4275: imp module is deprecated URL: https://github.com/apache/incubator-tvm/pull/4275#discussion_r346697376 ## File path: topi/python/topi/cpp/__init__.py ## @@ -0,0 +1,9 @@ +"""FFI for C++ TOPI ops and schedules"""

[GitHub] [incubator-tvm] t-vi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

t-vi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554248489 I changed the default to 256. This patch is independent of #4305 in terms of merge order (the decoupling of the device

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

were commented on issue #4275: imp module is deprecated URL: https://github.com/apache/incubator-tvm/pull/4275#issuecomment-554248004 @Laurawly @Huyuwei Any suggestions? This is an automated message from the Apache Git

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

were commented on issue #4275: imp module is deprecated URL: https://github.com/apache/incubator-tvm/pull/4275#issuecomment-554247918 In `topi.cpp`, a bunch of dummy modules are created to get rid of a bunch of modules with only one line.

[GitHub] [incubator-tvm] t-vi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen

t-vi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#discussion_r346690529 ## File path: src/codegen/llvm/codegen_amdgpu.cc ## @@ -36,13 +36,39 @@ namespace

[GitHub] [incubator-tvm] t-vi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen

t-vi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#discussion_r346690529 ## File path: src/codegen/llvm/codegen_amdgpu.cc ## @@ -36,13 +36,39 @@ namespace

[GitHub] [incubator-tvm] tmoreau89 commented on issue #4293: [VTA] Bug fix for padded load with large inputs

tmoreau89 commented on issue #4293: [VTA] Bug fix for padded load with large inputs URL: https://github.com/apache/incubator-tvm/pull/4293#issuecomment-554243956 Thank you @liangfu; if you could test the unit test on Pynq that would be excellent; one known issue is that the padded load

[GitHub] [incubator-tvm] liangfu commented on issue #4293: [VTA] Bug fix for padded load with large inputs

liangfu commented on issue #4293: [VTA] Bug fix for padded load with large inputs URL: https://github.com/apache/incubator-tvm/pull/4293#issuecomment-554241928 @vegaluisjose Yes, current Chisel implement passes all the unit-tests. @tmoreau89 I did not test this with HLS version yet.

[GitHub] [incubator-tvm] tmoreau89 commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose output_padding parameter

tmoreau89 commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose output_padding parameter URL: https://github.com/apache/incubator-tvm/pull/4318#issuecomment-554237615 I see; I'm a little confused because the test case with output padding set to `(0,0)` should be

[GitHub] [incubator-tvm] tmoreau89 closed pull request #3881: [CMAKE][VTA] Modularizing Runtime Sources

tmoreau89 closed pull request #3881: [CMAKE][VTA] Modularizing Runtime Sources URL: https://github.com/apache/incubator-tvm/pull/3881 This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [incubator-tvm] tmoreau89 commented on issue #3881: [CMAKE][VTA] Modularizing Runtime Sources

tmoreau89 commented on issue #3881: [CMAKE][VTA] Modularizing Runtime Sources URL: https://github.com/apache/incubator-tvm/pull/3881#issuecomment-554236128 Closing the PR due to minimal impact on compilation time. This is an

[GitHub] [incubator-tvm] abergeron commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose output_padding parameter

abergeron commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose output_padding parameter URL: https://github.com/apache/incubator-tvm/pull/4318#issuecomment-554228705 It fails with (0, 0) in the CI right now. It might also be a good idea to test it with non-zero

[GitHub] [incubator-tvm] wweic opened a new pull request #4346: [Runtime] Make ADTObject POD container type

wweic opened a new pull request #4346: [Runtime] Make ADTObject POD container type URL: https://github.com/apache/incubator-tvm/pull/4346 This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [incubator-tvm] masahi commented on issue #4305: [RUNTIME] Proper Device Attribute Query for AMD GPU

masahi commented on issue #4305: [RUNTIME] Proper Device Attribute Query for AMD GPU URL: https://github.com/apache/incubator-tvm/pull/4305#issuecomment-554211746 @petrex sorry I merged #4341 first and there is a conflict. Can you rebase?

[GitHub] [incubator-tvm] tqchen opened a new pull request #4345: [COMMUNITY] Add DISCLAIMER, KEYS for ASF release

tqchen opened a new pull request #4345: [COMMUNITY] Add DISCLAIMER, KEYS for ASF release URL: https://github.com/apache/incubator-tvm/pull/4345 cc @yzhliu This is an automated message from the Apache Git Service. To respond

[GitHub] [incubator-tvm] masahi merged pull request #4341: [RUNTIME] Add device query for AMD GcnArch

masahi merged pull request #4341: [RUNTIME] Add device query for AMD GcnArch URL: https://github.com/apache/incubator-tvm/pull/4341 This is an automated message from the Apache Git Service. To respond to the message, please

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#discussion_r346657735 ## File path: topi/python/topi/transform.py ## @@ -155,6 +157,78 @@ def strided_slice(a,

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#discussion_r346657674 ## File path: topi/python/topi/transform.py ## @@ -155,6 +157,78 @@ def strided_slice(a,

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#discussion_r346657709 ## File path: topi/python/topi/transform.py ## @@ -155,6 +157,78 @@ def strided_slice(a,

[GitHub] [incubator-tvm] junrushao1994 commented on issue #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on issue #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#issuecomment-554205813 @yzhliu I think this is something like `a[strides] = b`

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#discussion_r346656799 ## File path: python/tvm/relay/op/_transform.py ## @@ -304,6 +305,31 @@ def

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#discussion_r346657038 ## File path: python/tvm/relay/op/_transform.py ## @@ -304,6 +305,31 @@ def

[GitHub] [incubator-tvm] junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation.

junrushao1994 commented on a change in pull request #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#discussion_r346657023 ## File path: python/tvm/relay/op/_transform.py ## @@ -304,6 +305,31 @@ def

[GitHub] [incubator-tvm] FrozenGene opened a new pull request #4344: [ThreadPool] Solve thread transitions issue

FrozenGene opened a new pull request #4344: [ThreadPool] Solve thread transitions issue URL: https://github.com/apache/incubator-tvm/pull/4344 This pr solved thread transitions issue. When we use OpenCV + TVM, we will occur much thread transitions so that OpenCV + TVM is slow.

[GitHub] [incubator-tvm] tqchen commented on issue #4327: [Relay][Frontend][TF] Fix transpose when axes is not a param

tqchen commented on issue #4327: [Relay][Frontend][TF] Fix transpose when axes is not a param URL: https://github.com/apache/incubator-tvm/pull/4327#issuecomment-554200646 Thanks @soiferj This is an automated message from

[incubator-tvm] branch master updated (72821b2 -> 1e2c525)

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 72821b2 [Contrib] Add MKL DNN option (#4323) add 1e2c525 [Relay][Frontend][TF] Fix transpose when axes

[GitHub] [incubator-tvm] tqchen merged pull request #4327: [Relay][Frontend][TF] Fix transpose when axes is not a param

tqchen merged pull request #4327: [Relay][Frontend][TF] Fix transpose when axes is not a param URL: https://github.com/apache/incubator-tvm/pull/4327 This is an automated message from the Apache Git Service. To respond to

[GitHub] [incubator-tvm] tqchen closed pull request #3828: [QUANTIZE] Improve explicitness of rules during annotation/realization

tqchen closed pull request #3828: [QUANTIZE] Improve explicitness of rules during annotation/realization URL: https://github.com/apache/incubator-tvm/pull/3828 This is an automated message from the Apache Git Service. To

[GitHub] [incubator-tvm] tqchen commented on issue #3966: Const loop partition fix

tqchen commented on issue #3966: Const loop partition fix URL: https://github.com/apache/incubator-tvm/pull/3966#issuecomment-554199981 superceded by #4025 This is an automated message from the Apache Git Service. To respond

[GitHub] [incubator-tvm] tqchen closed pull request #3658: [Relay][MxNet] Add support for foreach

tqchen closed pull request #3658: [Relay][MxNet] Add support for foreach URL: https://github.com/apache/incubator-tvm/pull/3658 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [incubator-tvm] tqchen closed pull request #3966: Const loop partition fix

tqchen closed pull request #3966: Const loop partition fix URL: https://github.com/apache/incubator-tvm/pull/3966 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[GitHub] [incubator-tvm] tqchen edited a comment on issue #3966: Const loop partition fix

tqchen edited a comment on issue #3966: Const loop partition fix URL: https://github.com/apache/incubator-tvm/pull/3966#issuecomment-554199981 superceded by #4025 thanks @kimishpatel @yinghai This is an automated message

[GitHub] [incubator-tvm] tqchen commented on issue #3881: [CMAKE][VTA] Modularizing Runtime Sources

tqchen commented on issue #3881: [CMAKE][VTA] Modularizing Runtime Sources URL: https://github.com/apache/incubator-tvm/pull/3881#issuecomment-554199820 ping @tmoreau89 This is an automated message from the Apache Git

[GitHub] [incubator-tvm] tqchen edited a comment on issue #4323: [Contrib] Add MKL DNN option

tqchen edited a comment on issue #4323: [Contrib] Add MKL DNN option URL: https://github.com/apache/incubator-tvm/pull/4323#issuecomment-554199391 Thanks @icemelon9 @TaoLv @soiferj @gasgallo @minminsun ! This is an automated

[GitHub] [incubator-tvm] tqchen commented on issue #4323: [Contrib] Add MKL DNN option

tqchen commented on issue #4323: [Contrib] Add MKL DNN option URL: https://github.com/apache/incubator-tvm/pull/4323#issuecomment-554199391 Thanks @icemelon9 @TaoLv @soiferj ! This is an automated message from the Apache Git

[GitHub] [incubator-tvm] tqchen edited a comment on issue #4323: [Contrib] Add MKL DNN option

tqchen edited a comment on issue #4323: [Contrib] Add MKL DNN option URL: https://github.com/apache/incubator-tvm/pull/4323#issuecomment-554199391 Thanks @icemelon9 @TaoLv @soiferj @gasgallo @minminsun @ZhennanQin ! This is

[GitHub] [incubator-tvm] tqchen merged pull request #4333: Deprecate NNVM warning msg

tqchen merged pull request #4333: Deprecate NNVM warning msg URL: https://github.com/apache/incubator-tvm/pull/4333 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[incubator-tvm] branch master updated (de1bfa4 -> 2573b3b)

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 de1bfa4 Solve custom model of prelu (#4326) add 2573b3b Deprecate NNVM warning msg (#4333) No new

[GitHub] [incubator-tvm] tqchen commented on issue #4303: [TOPI][Relay][OP] Add a strided_set operation.

tqchen commented on issue #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#issuecomment-554199080 cc @jroesch @junrushao1994 can you also take a look? This is an

[GitHub] [incubator-tvm] tqchen merged pull request #4326: [TFLite] Fix Prelu unified shape error

tqchen merged pull request #4326: [TFLite] Fix Prelu unified shape error URL: https://github.com/apache/incubator-tvm/pull/4326 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [incubator-tvm] tqchen commented on issue #4326: [TFLite] Fix Prelu unified shape error

tqchen commented on issue #4326: [TFLite] Fix Prelu unified shape error URL: https://github.com/apache/incubator-tvm/pull/4326#issuecomment-554198952 Thanks @FrozenGene @apivovarov This is an automated message from the

[incubator-tvm] branch master updated (23691b8 -> de1bfa4)

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 23691b8 Add topi.nn.fifo_buffer to TVM doc (#4343) add de1bfa4 Solve custom model of prelu (#4326) No

[GitHub] [incubator-tvm] tqchen commented on issue #4343: Add topi.nn.fifo_buffer to TOPI API doc

tqchen commented on issue #4343: Add topi.nn.fifo_buffer to TOPI API doc URL: https://github.com/apache/incubator-tvm/pull/4343#issuecomment-554198837 Thanks @hcho3 @zhiics @yongwww @junrushao1994 This is an automated

[incubator-tvm] branch master updated (70a3a61 -> 23691b8)

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 70a3a61 Add support for quant. mul operator in tflite frontend (#4283) add 23691b8 Add

[GitHub] [incubator-tvm] tqchen merged pull request #4343: Add topi.nn.fifo_buffer to TOPI API doc

tqchen merged pull request #4343: Add topi.nn.fifo_buffer to TOPI API doc URL: https://github.com/apache/incubator-tvm/pull/4343 This is an automated message from the Apache Git Service. To respond to the message, please log

[incubator-tvm] branch master updated (7dca655 -> 70a3a61)

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 7dca655 [Relay][Pass] Add pass to remove unused functions in relay module (#4334) add 70a3a61 Add

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

tqchen commented on issue #4283: Add support for quant. mul operator in tflite frontend URL: https://github.com/apache/incubator-tvm/pull/4283#issuecomment-554198542 Thanks @u99127 @inadob @cchung100m This is an automated

[GitHub] [incubator-tvm] tqchen commented on a change in pull request #4274: [µTVM] Enable AutoTVM for ARM STM32F746XX Boards

tqchen commented on a change in pull request #4274: [µTVM] Enable AutoTVM for ARM STM32F746XX Boards URL: https://github.com/apache/incubator-tvm/pull/4274#discussion_r346651119 ## File path: src/runtime/micro/micro_session.cc ## @@ -76,79 +214,122 @@

[GitHub] [incubator-tvm] tqchen commented on issue #4340: [CodeGen] Add build config option disable_assert to control whether to generate assert

tqchen commented on issue #4340: [CodeGen] Add build config option disable_assert to control whether to generate assert URL: https://github.com/apache/incubator-tvm/pull/4340#issuecomment-554196234 We can do it one step before the final codegen

[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-554184646 LGTM! Thanks for putting up with all my feedback. This turned out to a be great PR. I think if you rebase

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

jwfromm removed a comment on issue #4271: [Relay][Frontend][ONNX] operator support: DepthToSpace, SpaceToDepth URL: https://github.com/apache/incubator-tvm/pull/4271#issuecomment-554184231 @cchung100m, now that PR #4313 is merged we should be able to get this PR merged. Can you take a

[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-554184231 @cchung100m, now that PR #4313 is merged we should be able to get this PR merged. Can you take a look at my

[GitHub] [incubator-tvm] FrozenGene commented on issue #4340: [CodeGen] Add build config option disable_assert to control whether to generate assert

FrozenGene commented on issue #4340: [CodeGen] Add build config option disable_assert to control whether to generate assert URL: https://github.com/apache/incubator-tvm/pull/4340#issuecomment-554183342 > Actually, it would be great if we can do it at an earlier point, and add a pass to

[GitHub] [incubator-tvm] vinx13 commented on issue #4295: [Relay][Quantize] Integrate data-aware calibration into quantization

vinx13 commented on issue #4295: [Relay][Quantize] Integrate data-aware calibration into quantization URL: https://github.com/apache/incubator-tvm/pull/4295#issuecomment-554178653 It's difficult to do unit tests. The refactor is covered by nightly tests. I plan to add more nightly tests

[GitHub] [incubator-tvm] tqchen commented on issue #4335: [Relay][WIP] ConvertLayout pass.

tqchen commented on issue #4335: [Relay][WIP] ConvertLayout pass. URL: https://github.com/apache/incubator-tvm/pull/4335#issuecomment-554175946 I feel live patching the OpAttr is not a pattern we are looking for in here. It would be great if we can factor out some common logic between

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

tqchen commented on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-554174983 @ZihengJiang can you help look into update the benchmark script? This is

[GitHub] [incubator-tvm] zhiics edited a comment on issue #4334: [Relay][Pass] Add pass to remove unused functions in relay module

zhiics edited a comment on issue #4334: [Relay][Pass] Add pass to remove unused functions in relay module URL: https://github.com/apache/incubator-tvm/pull/4334#issuecomment-554173088 Thanks @wweic @icemelon9 @MarisaKirisame

[GitHub] [incubator-tvm] zhiics commented on issue #4334: [Relay][Pass] Add pass to remove unused functions in relay module

zhiics commented on issue #4334: [Relay][Pass] Add pass to remove unused functions in relay module URL: https://github.com/apache/incubator-tvm/pull/4334#issuecomment-554173088 Thanks @wweic @jroesch @MarisaKirisame This is

[GitHub] [incubator-tvm] jackwish commented on issue #4338: AutoTVM: selecting tuning templates when extracting task

jackwish commented on issue #4338: AutoTVM: selecting tuning templates when extracting task URL: https://github.com/apache/incubator-tvm/pull/4338#issuecomment-554173072 Thank you for the review @comaniac @tmoreau89 @vinx13 , really good points to have a dict holding the op to key

[incubator-tvm] branch master updated (5b9f459 -> 7dca655)

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 5b9f459 Enable hipModuleGetGlobal() (#4321) add 7dca655 [Relay][Pass] Add pass to remove unused

[GitHub] [incubator-tvm] zhiics merged pull request #4334: [Relay][Pass] Add pass to remove unused functions in relay module

zhiics merged pull request #4334: [Relay][Pass] Add pass to remove unused functions in relay module URL: https://github.com/apache/incubator-tvm/pull/4334 This is an automated message from the Apache Git Service. To respond

[GitHub] [incubator-tvm] jackwish commented on a change in pull request #4338: AutoTVM: selecting tuning templates when extracting task

jackwish commented on a change in pull request #4338: AutoTVM: selecting tuning templates when extracting task URL: https://github.com/apache/incubator-tvm/pull/4338#discussion_r346632104 ## File path: python/tvm/autotvm/task/relay_integration.py ## @@ -98,6 +103,8 @@ def

[GitHub] [incubator-tvm] TaoLv commented on a change in pull request #4323: [Contrib] Add MKL DNN option

TaoLv commented on a change in pull request #4323: [Contrib] Add MKL DNN option URL: https://github.com/apache/incubator-tvm/pull/4323#discussion_r346631647 ## File path: cmake/modules/contrib/BLAS.cmake ## @@ -55,3 +55,11 @@ elseif(USE_BLAS STREQUAL "none") else()

[GitHub] [incubator-tvm] optima2005 commented on issue #4300: [Relay][Frontend][Tensorflow]Add conv2d_transpose

optima2005 commented on issue #4300: [Relay][Frontend][Tensorflow]Add conv2d_transpose URL: https://github.com/apache/incubator-tvm/pull/4300#issuecomment-554170862 tf1.4 doesn't have this error while tf1.3 does. This is an

[GitHub] [incubator-tvm] ZihengJiang edited a comment on issue #4332: [RFC] Support for Sparse Computation

ZihengJiang edited a comment on issue #4332: [RFC] Support for Sparse Computation URL: https://github.com/apache/incubator-tvm/issues/4332#issuecomment-553681206 Welcome comment and discussion! @cylinbao @yuluny2 @tmoreau89 @Huyuwei @yzh119 @tqchen

[GitHub] [incubator-tvm] szha commented on issue #4332: [RFC] Support for Sparse Computation

szha commented on issue #4332: [RFC] Support for Sparse Computation URL: https://github.com/apache/incubator-tvm/issues/4332#issuecomment-554159846 cc @eric-haibin-lin @sxjscience This is an automated message from the Apache

[GitHub] [incubator-tvm] anijain2305 commented on issue #4292: Retain qnn input kernel scales

anijain2305 commented on issue #4292: Retain qnn input kernel scales URL: https://github.com/apache/incubator-tvm/pull/4292#issuecomment-554159517 https://github.com/apache/incubator-tvm/blob/5b9f459d638d1cf6bd820f3bdc58d7e5632d7ed7/python/tvm/relay/qnn/op/legalizations.py#L89-L90

[GitHub] [incubator-tvm] cylinbao commented on issue #4332: [RFC] Support for Sparse Computation

cylinbao commented on issue #4332: [RFC] Support for Sparse Computation URL: https://github.com/apache/incubator-tvm/issues/4332#issuecomment-554158427 Thank @ZihengJiang for bringing up the RFC. One question. It seems like, in this RFC, the axis to store the data values is always the

[GitHub] [incubator-tvm] u99127 commented on issue #4292: Retain qnn input kernel scales

u99127 commented on issue #4292: Retain qnn input kernel scales URL: https://github.com/apache/incubator-tvm/pull/4292#issuecomment-554157700 Ok, it looks like there is some more work to be done here because of other bits that have landed. - Fix legalisations.py to consider qnn.conv2d

[GitHub] [incubator-tvm] masahi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen

masahi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#discussion_r346617474 ## File path: src/codegen/llvm/codegen_amdgpu.cc ## @@ -36,13 +36,39 @@ namespace

[GitHub] [incubator-tvm] masahi merged pull request #4321: [RUNTIME] Enable hipModuleGetGlobal() in ROCm module

masahi merged pull request #4321: [RUNTIME] Enable hipModuleGetGlobal() in ROCm module URL: https://github.com/apache/incubator-tvm/pull/4321 This is an automated message from the Apache Git Service. To respond to the

[incubator-tvm] branch master updated (d9b8a6c -> 5b9f459)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from d9b8a6c [Build][Windows] Fix Windows build by including cctype (#4319) add 5b9f459 Enable

[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_r346612067 ## File path: tests/python/frontend/onnx/test_forward.py ## @@ -77,22

[GitHub] [incubator-tvm] u99127 commented on issue #4320: Fix TFLite RESHAPE assert

u99127 commented on issue #4320: Fix TFLite RESHAPE assert URL: https://github.com/apache/incubator-tvm/pull/4320#issuecomment-554135038 I've been noticing differences in the tflite converter graph between 1.13 and 1.14 especially with split and unpack - have either of you guys played with

[GitHub] [incubator-tvm] t-vi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen

t-vi commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#discussion_r346595415 ## File path: src/codegen/llvm/codegen_amdgpu.cc ## @@ -36,13 +36,39 @@ namespace

[incubator-tvm] branch master updated (5d66e7a -> d9b8a6c)

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 5d66e7a [CI] Set workspace to be per executor (#4336) add d9b8a6c [Build][Windows] Fix Windows build

[GitHub] [incubator-tvm] yzhliu commented on issue #4319: [Build][Windows] Fix Windows build by including cctype

yzhliu commented on issue #4319: [Build][Windows] Fix Windows build by including cctype URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-554128110 Thanks @soiferj @jmorrill @tqchen This is an automated

[GitHub] [incubator-tvm] yzhliu merged pull request #4319: [Build][Windows] Fix Windows build by including cctype

yzhliu merged pull request #4319: [Build][Windows] Fix Windows build by including cctype URL: https://github.com/apache/incubator-tvm/pull/4319 This is an automated message from the Apache Git Service. To respond to the

[GitHub] [incubator-tvm] zhiics commented on issue #4179: Add More Shape Functions

zhiics commented on issue #4179: Add More Shape Functions URL: https://github.com/apache/incubator-tvm/pull/4179#issuecomment-554127427 I would prefer register_rt_shape_func or just keep register_shape_func. Is it bad to think it is similar to type_rel?

[GitHub] [incubator-tvm] petrex commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen

petrex commented on a change in pull request #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#discussion_r346592319 ## File path: src/codegen/llvm/codegen_amdgpu.cc ## @@ -36,13 +36,39 @@ namespace

[GitHub] [incubator-tvm] icemelon9 commented on issue #4179: Add More Shape Functions

icemelon9 commented on issue #4179: Add More Shape Functions URL: https://github.com/apache/incubator-tvm/pull/4179#issuecomment-554123136 JIT is not accurate here since these shape function are compiled at compilation time instead of runtime. We can probably use `register_rt_shape_func`

[GitHub] [incubator-tvm] apivovarov commented on issue #4326: [TFLite] Fix Prelu unified shape error

apivovarov commented on issue #4326: [TFLite] Fix Prelu unified shape error URL: https://github.com/apache/incubator-tvm/pull/4326#issuecomment-554122057 +1 This is an automated message from the Apache Git Service. To respond

[GitHub] [incubator-tvm] ZihengJiang commented on issue #4243: Fix broken loop partitioning due to recent changes.

ZihengJiang commented on issue #4243: Fix broken loop partitioning due to recent changes. URL: https://github.com/apache/incubator-tvm/pull/4243#issuecomment-554121953 Approved. Thanks for contributing! @kimishpatel This is

[GitHub] [incubator-tvm] yzhliu commented on issue #4315: Add check to ensure input file was successfully opened in NNVM deploy…

yzhliu commented on issue #4315: Add check to ensure input file was successfully opened in NNVM deploy… URL: https://github.com/apache/incubator-tvm/pull/4315#issuecomment-554115882 please rebase to retrigger the ci This is

[GitHub] [incubator-tvm] masahi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

masahi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554114635 Great! Thanks. This is an automated message from the

[GitHub] [incubator-tvm] petrex commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

petrex commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554114502 Sure, Let's not used hardcoded value. btw you are testing on gfx900 right (or another arch)?

[GitHub] [incubator-tvm] petrex edited a comment on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

petrex edited a comment on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554114502 Sure, Let's not use hardcoded value. btw you are testing on gfx900 right (or another arch)?

[GitHub] [incubator-tvm] t-vi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

t-vi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554111798 @masahi Yes, indeed, it fixes this, in fact I'm doing this with @mvermeulen 's tests in mind. @petrex So this

[GitHub] [incubator-tvm] yzhliu commented on issue #4303: [TOPI][Relay][OP] Add a strided_set operation.

yzhliu commented on issue #4303: [TOPI][Relay][OP] Add a strided_set operation. URL: https://github.com/apache/incubator-tvm/pull/4303#issuecomment-554109627 is it used in dl frameworks? This is an automated message from the

[GitHub] [incubator-tvm] masahi edited a comment on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

masahi edited a comment on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554107796 does this solve the issue of INVALID_ISA errors[ discussed in the forum

[GitHub] [incubator-tvm] masahi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

masahi commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554107796 does this solve the issue of INVALID_ISA errors[ discussed in the forum

[GitHub] [incubator-tvm] comaniac commented on issue #4280: [TVM][RUNTIME] A minimum example to generate external library wrappers for DSOModule

comaniac commented on issue #4280: [TVM][RUNTIME] A minimum example to generate external library wrappers for DSOModule URL: https://github.com/apache/incubator-tvm/pull/4280#issuecomment-554107934 cc @tqchen This is an

[GitHub] [incubator-tvm] tmoreau89 commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose output_padding parameter

tmoreau89 commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose output_padding parameter URL: https://github.com/apache/incubator-tvm/pull/4318#issuecomment-554107152 @abergeron is the issue right now that VTA test will fail when you set output padding to `(1,1)`? Or does

[GitHub] [incubator-tvm] petrex commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen

petrex commented on issue #4342: Add workgroup size attribute to AMDGPU functions in codegen URL: https://github.com/apache/incubator-tvm/pull/4342#issuecomment-554106244 Thanks @t-vi . @masahi This PR utilizes device query for `kMaxThreadsPerBlock`, in that case we might need to

[GitHub] [incubator-tvm] hcho3 opened a new pull request #4343: Add topi.nn.fifo_buffer to TOPI API doc

hcho3 opened a new pull request #4343: Add topi.nn.fifo_buffer to TOPI API doc URL: https://github.com/apache/incubator-tvm/pull/4343 Follow-up to #4039. Add `topi.nn.fifo_buffer` to the TOPI API doc. This is an automated

[GitHub] [incubator-tvm] apivovarov edited a comment on issue #4320: Fix TFLite RESHAPE assert

apivovarov edited a comment on issue #4320: Fix TFLite RESHAPE assert URL: https://github.com/apache/incubator-tvm/pull/4320#issuecomment-554081798 TVM code does not use the second input tensor in RESHAPE op. From the beginning the assert should just check that input tensors array length

[GitHub] [incubator-tvm] wweic commented on a change in pull request #4334: [Relay][Pass] Add pass to remove unused functions in relay module

wweic commented on a change in pull request #4334: [Relay][Pass] Add pass to remove unused functions in relay module URL: https://github.com/apache/incubator-tvm/pull/4334#discussion_r346548590 ## File path: src/relay/backend/vm/compiler.cc ## @@ -863,6 +864,8 @@ void

[GitHub] [incubator-tvm] apivovarov commented on issue #4320: Fix TFLite RESHAPE assert

apivovarov commented on issue #4320: Fix TFLite RESHAPE assert URL: https://github.com/apache/incubator-tvm/pull/4320#issuecomment-554081798 TVM code does not use the second input tensor in RESHAPE op. From the beginning the assert should just check that input tensors array length is at

[GitHub] [incubator-tvm] yzhliu commented on a change in pull request #4333: Deprecate NNVM warning msg

yzhliu commented on a change in pull request #4333: Deprecate NNVM warning msg URL: https://github.com/apache/incubator-tvm/pull/4333#discussion_r346547323 ## File path: nnvm/python/nnvm/__init__.py ## @@ -10,3 +10,4 @@ from . import frontend __version__ =

  1   2   >