[GitHub] [incubator-tvm] yzhliu commented on issue #5121: [TE] reverse-mode autodiff without any optimization

yzhliu commented on issue #5121: [TE] reverse-mode autodiff without any optimization URL: https://github.com/apache/incubator-tvm/pull/5121#issuecomment-603648759 @MarisaKirisame sure, I will try. This is an automated

[GitHub] [incubator-tvm] ANSHUMAN87 commented on issue #5084: Duplicate likely nodes added when loop axis split unevenly

ANSHUMAN87 commented on issue #5084: Duplicate likely nodes added when loop axis split unevenly URL: https://github.com/apache/incubator-tvm/pull/5084#issuecomment-603636335 @yzhliu: Gentle reminder! This is an automated

[GitHub] [incubator-tvm] tqchen commented on issue #5060: [uTVM][Runtime] Deprecate uTVM Standalone Runtime

tqchen commented on issue #5060: [uTVM][Runtime] Deprecate uTVM Standalone Runtime URL: https://github.com/apache/incubator-tvm/issues/5060#issuecomment-603632782 re: fragmentation issue, think the allocation strategies carefully and adopt an arena-style allocator(counter based as above)

[GitHub] [incubator-tvm] MarisaKirisame commented on a change in pull request #5144: [Relay][VM] Memory planner (part 1)

MarisaKirisame commented on a change in pull request #5144: [Relay][VM] Memory planner (part 1) URL: https://github.com/apache/incubator-tvm/pull/5144#discussion_r397530214 ## File path: python/tvm/relay/transform/memory_plan.py ## @@ -0,0 +1,189 @@ +# Licensed to the

[incubator-tvm] branch master updated (686911e -> 3aabbd9)

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 686911e [Torch] Fix conv2d conversion for group conv (group > 1 but != in channels) (#5132) add

[GitHub] [incubator-tvm] masahi merged pull request #5142: [Torch] Add support for max_pool1d

masahi merged pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142 This is an automated message from the Apache Git Service. To respond to the message, please log on

[GitHub] [incubator-tvm] masahi commented on issue #5142: [Torch] Add support for max_pool1d

masahi commented on issue #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#issuecomment-603510415 Thanks @wyc-ruiker This is an automated message from the Apache Git Service.

[GitHub] [incubator-tvm] masahi commented on issue #5134: [RELAY] Add MergeCompilerRegions pass

masahi commented on issue #5134: [RELAY] Add MergeCompilerRegions pass URL: https://github.com/apache/incubator-tvm/pull/5134#issuecomment-603507803 @mbaret Sorry I haven't been following the development of a new partitioning algorithm. Is this related to annotator support for composite?

[GitHub] [incubator-tvm] jroesch commented on issue #5144: [Relay][VM] Memory planner (part 1)

jroesch commented on issue #5144: [Relay][VM] Memory planner (part 1) URL: https://github.com/apache/incubator-tvm/pull/5144#issuecomment-603496275 cc @zhiics @yidawang @icemelon9 @MarisaKirisame This is an automated message

[GitHub] [incubator-tvm] jroesch opened a new pull request #5144: [Relay][VM] Memory planner (part 1)

jroesch opened a new pull request #5144: [Relay][VM] Memory planner (part 1) URL: https://github.com/apache/incubator-tvm/pull/5144 This PR adds a new pass to the memory scheduling phase of the VM. After we have made all memory allocations explicit, we now analyze for basic block style

[GitHub] [incubator-tvm] tqchen commented on issue #5132: [Torch] Fix conv2d conversion for group conv (group > 1 but != in channels)

tqchen commented on issue #5132: [Torch] Fix conv2d conversion for group conv (group > 1 but != in channels) URL: https://github.com/apache/incubator-tvm/pull/5132#issuecomment-603470964 Thanks @masahi @jwfromm ! This is an

[GitHub] [incubator-tvm] MarisaKirisame commented on issue #5121: [TE] reverse-mode autodiff without any optimization

MarisaKirisame commented on issue #5121: [TE] reverse-mode autodiff without any optimization URL: https://github.com/apache/incubator-tvm/pull/5121#issuecomment-603454210 (of course, not in this PR) This is an automated

[GitHub] [incubator-tvm] MarisaKirisame commented on issue #5121: [TE] reverse-mode autodiff without any optimization

MarisaKirisame commented on issue #5121: [TE] reverse-mode autodiff without any optimization URL: https://github.com/apache/incubator-tvm/pull/5121#issuecomment-603454001 @yzhliu can you do forward mode automatic differentiation? It is easy considering you have jacobian - you only need to

[GitHub] [incubator-tvm] yzhliu commented on issue #5121: [TE] reverse-mode autodiff without any optimization

yzhliu commented on issue #5121: [TE] reverse-mode autodiff without any optimization URL: https://github.com/apache/incubator-tvm/pull/5121#issuecomment-603446436 @MarisaKirisame @tqchen @hzfan Could you review again? This

[GitHub] [incubator-tvm] hypercubestart commented on issue #5039: [Relay] GradientCell Relay Pass

hypercubestart commented on issue #5039: [Relay] GradientCell Relay Pass URL: https://github.com/apache/incubator-tvm/pull/5039#issuecomment-603437659 thanks @jroesch ! This is an automated message from the Apache Git

[GitHub] [incubator-tvm] icemelon9 commented on issue #5131: [Relay][TOPI] Register topi schedule for Relay fast_exp and fast_tanh

icemelon9 commented on issue #5131: [Relay][TOPI] Register topi schedule for Relay fast_exp and fast_tanh URL: https://github.com/apache/incubator-tvm/pull/5131#issuecomment-603427810 @selo1412 Could you address @comaniac comments?

[GitHub] [incubator-tvm] u99127 commented on issue #5060: [uTVM][Runtime] Deprecate uTVM Standalone Runtime

u99127 commented on issue #5060: [uTVM][Runtime] Deprecate uTVM Standalone Runtime URL: https://github.com/apache/incubator-tvm/issues/5060#issuecomment-603408836 Thanks for pointing this to me @tmoreau89 and thank you for this work @liangfu . Very interesting and good questions to ask.

[GitHub] [incubator-tvm] manupa-arm commented on issue #5143: [RELAY] Re-wrote the Graph Partitioner to support multiple outputs

manupa-arm commented on issue #5143: [RELAY] Re-wrote the Graph Partitioner to support multiple outputs URL: https://github.com/apache/incubator-tvm/pull/5143#issuecomment-603402680 cc : @zhiics @comaniac @tqchen This is an

[GitHub] [incubator-tvm] manupa-arm opened a new pull request #5143: [RELAY] Re-wrote the Graph Partitioner to support multiple outputs

manupa-arm opened a new pull request #5143: [RELAY] Re-wrote the Graph Partitioner to support multiple outputs URL: https://github.com/apache/incubator-tvm/pull/5143 This PR aims to **include support for multiple outputs** in the regions that get outlined as functions for different

[GitHub] [incubator-tvm] trevor-m commented on a change in pull request #5106: [Relay][MergeComposite] Support TupleGetItem in body of pattern

trevor-m commented on a change in pull request #5106: [Relay][MergeComposite] Support TupleGetItem in body of pattern URL: https://github.com/apache/incubator-tvm/pull/5106#discussion_r397337604 ## File path: src/relay/transforms/merge_composite.cc ## @@ -66,6 +66,31 @@

[incubator-tvm] branch master updated (0a0e58b -> 686911e)

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 0a0e58b [REFACTOR][TIR] Introduce PrimFuncPass. (#5139) add 686911e [Torch] Fix conv2d conversion for

[GitHub] [incubator-tvm] zhiics commented on a change in pull request #5028: [RELAY] Remove kCompiler attr from ext mod

zhiics commented on a change in pull request #5028: [RELAY] Remove kCompiler attr from ext mod URL: https://github.com/apache/incubator-tvm/pull/5028#discussion_r397321424 ## File path: src/relay/backend/compile_engine.cc ## @@ -627,6 +627,7 @@ class CompileEngineImpl :

[GitHub] [incubator-tvm] tqchen merged pull request #5132: [Torch] Fix conv2d conversion for group conv (group > 1 but != in channels)

tqchen merged pull request #5132: [Torch] Fix conv2d conversion for group conv (group > 1 but != in channels) URL: https://github.com/apache/incubator-tvm/pull/5132 This is an automated message from the Apache Git Service.

[GitHub] [incubator-tvm] tqchen commented on issue #5060: [uTVM][Runtime] Deprecate uTVM Standalone Runtime

tqchen commented on issue #5060: [uTVM][Runtime] Deprecate uTVM Standalone Runtime URL: https://github.com/apache/incubator-tvm/issues/5060#issuecomment-603374417 The workspace memory could have a different strategy. The way it works is that we create a different arena for workspace,

[GitHub] [incubator-tvm] zhiics commented on a change in pull request #5028: [RELAY] Remove kCompiler attr from ext mod

zhiics commented on a change in pull request #5028: [RELAY] Remove kCompiler attr from ext mod URL: https://github.com/apache/incubator-tvm/pull/5028#discussion_r397314736 ## File path: src/relay/backend/compile_engine.cc ## @@ -627,6 +627,7 @@ class CompileEngineImpl :

[GitHub] [incubator-tvm] trevor-m commented on issue #4847: Return empty CSourceModule when no lowered_funcs exists in Relay mod

trevor-m commented on issue #4847: Return empty CSourceModule when no lowered_funcs exists in Relay mod URL: https://github.com/apache/incubator-tvm/pull/4847#issuecomment-603369453 > I think changing it to a llvm module and import all submodules is okay. Now if you only have an external

[GitHub] [incubator-tvm] mehrdadhe commented on issue #3934: [Runtime] MISRA-C compliant TVM runtime

mehrdadhe commented on issue #3934: [Runtime] MISRA-C compliant TVM runtime URL: https://github.com/apache/incubator-tvm/pull/3934#issuecomment-603346577 Great work. @liangfu have you considered using "[system lib](https://docs.tvm.ai/api/python/target.html?highlight=system%20lib)"

[GitHub] [incubator-tvm] zhiics commented on a change in pull request #5030: [RELAY] Added a AnnotatedRegion utility class

zhiics commented on a change in pull request #5030: [RELAY] Added a AnnotatedRegion utility class URL: https://github.com/apache/incubator-tvm/pull/5030#discussion_r397261539 ## File path: python/tvm/relay/analysis/annotated_regions.py ## @@ -0,0 +1,62 @@ +# Licensed to

[GitHub] [incubator-tvm] tqchen edited a comment on issue #5124: [uTVM][Runtime] Introduce Virtual Memory Allocator to CRT

tqchen edited a comment on issue #5124: [uTVM][Runtime] Introduce Virtual Memory Allocator to CRT URL: https://github.com/apache/incubator-tvm/pull/5124#issuecomment-603312590 The workspace memory could have a different strategy. The way it works is that we create a different arena for

[GitHub] [incubator-tvm] tqchen commented on issue #5124: [uTVM][Runtime] Introduce Virtual Memory Allocator to CRT

tqchen commented on issue #5124: [uTVM][Runtime] Introduce Virtual Memory Allocator to CRT URL: https://github.com/apache/incubator-tvm/pull/5124#issuecomment-603312590 The workspace memory could have a different strategy. The way it works is that we create a different arena for

[GitHub] [incubator-tvm] Shawn-Inspur commented on a change in pull request #5099: [TOPI][Tensor Core] Conv2d and Dense ops support on Tensor Core

Shawn-Inspur commented on a change in pull request #5099: [TOPI][Tensor Core] Conv2d and Dense ops support on Tensor Core URL: https://github.com/apache/incubator-tvm/pull/5099#discussion_r397237931 ## File path: topi/python/topi/cuda/conv2d_nhwc_tensorcore.py ## @@ -0,0

[incubator-tvm] branch master updated: [REFACTOR][TIR] Introduce PrimFuncPass. (#5139)

This is an automated email from the ASF dual-hosted git repository. tqchen 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 0a0e58b [REFACTOR][TIR] Introduce

[GitHub] [incubator-tvm] tqchen merged pull request #5139: [REFACTOR][TIR] Introduce PrimFuncPass.

tqchen merged pull request #5139: [REFACTOR][TIR] Introduce PrimFuncPass. URL: https://github.com/apache/incubator-tvm/pull/5139 This is an automated message from the Apache Git Service. To respond to the message, please log

[GitHub] [incubator-tvm] tqchen commented on issue #5139: [REFACTOR][TIR] Introduce PrimFuncPass.

tqchen commented on issue #5139: [REFACTOR][TIR] Introduce PrimFuncPass. URL: https://github.com/apache/incubator-tvm/pull/5139#issuecomment-603298988 Thanks @zhiics @siju-samue哈 This is an automated message from the Apache

[incubator-tvm] branch master updated: [RUNTIME]fix unused-value warning (#5140)

This is an automated email from the ASF dual-hosted git repository. tqchen 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 bbbfc1b [RUNTIME]fix unused-value

[GitHub] [incubator-tvm] tqchen merged pull request #5140: [RUNTIME]fix unused-value warning

tqchen merged pull request #5140: [RUNTIME]fix unused-value warning URL: https://github.com/apache/incubator-tvm/pull/5140 This is an automated message from the Apache Git Service. To respond to the message, please log on to

[GitHub] [incubator-tvm] wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d

wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397132458 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d

wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397132370 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d

wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397124944 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d

masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397118351 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #5106: [Relay][MergeComposite] Support TupleGetItem in body of pattern

mbaret commented on a change in pull request #5106: [Relay][MergeComposite] Support TupleGetItem in body of pattern URL: https://github.com/apache/incubator-tvm/pull/5106#discussion_r397085411 ## File path: src/relay/transforms/merge_composite.cc ## @@ -66,6 +66,31 @@

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #5106: [Relay][MergeComposite] Support TupleGetItem in body of pattern

mbaret commented on a change in pull request #5106: [Relay][MergeComposite] Support TupleGetItem in body of pattern URL: https://github.com/apache/incubator-tvm/pull/5106#discussion_r397086345 ## File path: src/relay/transforms/merge_composite.cc ## @@ -66,6 +66,31 @@

[GitHub] [incubator-tvm] wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d

wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397074587 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d

wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397074097 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d

wyc-ruiker commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397074097 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d

masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397050020 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d

masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397049896 ## File path: src/relay/op/nn/pooling.cc ## @@ -987,6 +987,10 @@ Array Pool1DCompute(const Attrs& attrs,

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d

masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397048418 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -213,12 +213,33 @@ def _impl(inputs, input_types):

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d

masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397048418 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -213,12 +213,33 @@ def _impl(inputs, input_types):

[GitHub] [incubator-tvm] masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d

masahi commented on a change in pull request #5142: [Torch] Add support for max_pool1d URL: https://github.com/apache/incubator-tvm/pull/5142#discussion_r397046667 ## File path: tests/python/frontend/pytorch/test_forward.py ## @@ -363,9 +363,35 @@ class

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #5030: [RELAY] Added a AnnotatedRegion utility class

mbaret commented on a change in pull request #5030: [RELAY] Added a AnnotatedRegion utility class URL: https://github.com/apache/incubator-tvm/pull/5030#discussion_r397040388 ## File path: src/relay/analysis/annotated_region_set.cc ## @@ -0,0 +1,239 @@ +/* + * Licensed to

[GitHub] [incubator-tvm] mbaret commented on a change in pull request #5030: [RELAY] Added a AnnotatedRegion utility class

mbaret commented on a change in pull request #5030: [RELAY] Added a AnnotatedRegion utility class URL: https://github.com/apache/incubator-tvm/pull/5030#discussion_r397019200 ## File path: python/tvm/relay/analysis/annotated_regions.py ## @@ -0,0 +1,62 @@ +# Licensed to

[incubator-tvm] branch master updated (e6dd8e1 -> 667e24f)

This is an automated email from the ASF dual-hosted git repository. jroesch pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from e6dd8e1 [Relay] GradientCell Relay Pass (#5039) add 667e24f [CI] Update rust docker (#5141) No new

[GitHub] [incubator-tvm] jroesch merged pull request #5141: [CI] Update rust docker

jroesch merged pull request #5141: [CI] Update rust docker URL: https://github.com/apache/incubator-tvm/pull/5141 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub

[GitHub] [incubator-tvm] yzhliu commented on a change in pull request #5121: [TE] reverse-mode autodiff without any optimization

yzhliu commented on a change in pull request #5121: [TE] reverse-mode autodiff without any optimization URL: https://github.com/apache/incubator-tvm/pull/5121#discussion_r396935671 ## File path: src/te/autodiff/jacobian.cc ## @@ -0,0 +1,381 @@ +/* + * Licensed to the

[GitHub] [incubator-tvm] yzhliu commented on a change in pull request #5121: [TE] reverse-mode autodiff without any optimization

yzhliu commented on a change in pull request #5121: [TE] reverse-mode autodiff without any optimization URL: https://github.com/apache/incubator-tvm/pull/5121#discussion_r396935671 ## File path: src/te/autodiff/jacobian.cc ## @@ -0,0 +1,381 @@ +/* + * Licensed to the

[incubator-tvm] branch master updated (a6de507 -> e6dd8e1)

This is an automated email from the ASF dual-hosted git repository. jroesch pushed a change to branch master in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. from a6de507 Change Rust version to stable in Docker (#5138) add e6dd8e1 [Relay] GradientCell Relay Pass

[GitHub] [incubator-tvm] jroesch merged pull request #5039: [Relay] GradientCell Relay Pass

jroesch merged pull request #5039: [Relay] GradientCell Relay Pass URL: https://github.com/apache/incubator-tvm/pull/5039 This is an automated message from the Apache Git Service. To respond to the message, please log on to