[GitHub] [tvm] jikechao commented on pull request #14604: [Array] Add deduction guide for Array(IterType,IterType) constructor

jikechao commented on PR #14604: URL: https://github.com/apache/tvm/pull/14604#issuecomment-1515766113 @Lunderberg Thanks for your efforts. Indeed, only one instance of this issue. After changing Array{...} to Array{...}. and changing the source code you provided [above](https://gi

[GitHub] [tvm] Hzfengsy commented on pull request #14492: [TIR] Flatten SeqStmt on construction

Hzfengsy commented on PR #14492: URL: https://github.com/apache/tvm/pull/14492#issuecomment-1515752734 Thanks @Lunderberg for the PR :) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specifi

[GitHub] [tvm] masahi commented on pull request #14676: [Community] Jiajun Jiang -> Reviewer

masahi commented on PR #14676: URL: https://github.com/apache/tvm/pull/14676#issuecomment-1515752151 @junrushao Need to resolve the conflict... -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the

[GitHub] [tvm] Hzfengsy merged pull request #14492: [TIR] Flatten SeqStmt on construction

Hzfengsy merged PR #14492: URL: https://github.com/apache/tvm/pull/14492 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.

[tvm] branch main updated (a8e3d75ec6 -> 2bb9698be9)

This is an automated email from the ASF dual-hosted git repository. syfeng pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from a8e3d75ec6 [Community] Qiang Zhang -> Reviewer (#14677) add 2bb9698be9 [TIR] Flatten SeqStmt on construction (#14492)

[tvm] branch main updated (7c97c4eb8d -> a8e3d75ec6)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from 7c97c4eb8d [Runtime] Fix Can't "query_imports" Bug of VM Executable (#14656) add a8e3d75ec6 [Community] Qiang Zhang -

[GitHub] [tvm] masahi merged pull request #14677: [Community] Qiang Zhang -> Reviewer

masahi merged PR #14677: URL: https://github.com/apache/tvm/pull/14677 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or

[GitHub] [tvm] jikechao commented on issue #14638: [Bug][ONNX] use `len()` to extract the length of `tvm.relay.expr.Var` leads to crash

jikechao commented on issue #14638: URL: https://github.com/apache/tvm/issues/14638#issuecomment-1515748488 Hello @vvchernov, thank for your comments. This is a reproducible script that only uses `relay.frontend.from_onnx()` API. ``` import onnx import tvm from tvm import re

[GitHub] [tvm] masahi commented on pull request #14675: [Unity] Use custom hash in `BlockBuilder` to avoid hashing large constants

masahi commented on PR #14675: URL: https://github.com/apache/tvm/pull/14675#issuecomment-1515746946 > can we try keep structural equality, and cache the structural hash(can ignore bind values)(and use that to check inequality? I haven't modified the equality part. Only the hash part

[GitHub] [tvm] tvm-bot commented on pull request #14678: [COLLAGE] Enable opencl adreno texture and CLML target

tvm-bot commented on PR #14678: URL: https://github.com/apache/tvm/pull/14678#issuecomment-1515724586 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] krishnaraj36 opened a new pull request, #14678: [COLLAGE] Enable opencl adreno texture and CLML target

krishnaraj36 opened a new pull request, #14678: URL: https://github.com/apache/tvm/pull/14678 Enable opencl texture and CLML targets for adreno in collage tuning with graph executor cost function. Few fix to support graph executor compilation. -- This is an automated message from the A

[tvm] branch nightly updated (8e4cedc4e3 -> 7c97c4eb8d)

This is an automated email from the ASF dual-hosted git repository. github-bot pushed a change to branch nightly in repository https://gitbox.apache.org/repos/asf/tvm.git from 8e4cedc4e3 [Docker] Support rootless docker when using docker/bash.sh (#14590) add fc7ca1f503 [tests][scripts]

[GitHub] [tvm] jikechao commented on issue #14652: [Bug][ONNX] CumSum crashed when call infer_value for axis

jikechao commented on issue #14652: URL: https://github.com/apache/tvm/issues/14652#issuecomment-1515670955 Thank you @vvchernov @ekalda, The axis is a 1-dim input of the model, thus the shape is not displayed in the graph. The detail onnx model info is: ![image](https://user-imag

[GitHub] [tvm] jikechao commented on issue #14661: [Bug][ONNX] cann't parse strides for LpPool

jikechao commented on issue #14661: URL: https://github.com/apache/tvm/issues/14661#issuecomment-1515644524 Hello, @AndrewZhaoLuo @vvchernov : It checked the source code, and Lppool is supported by onnx converter. I cannot figure out why `strides` attribute cannot be parsed. Is it my wron

[GitHub] [tvm] jikechao commented on issue #14660: [BUG][ONNX] indices_or_sections need to be able to divide input.shape[axis] in split op

jikechao commented on issue #14660: URL: https://github.com/apache/tvm/issues/14660#issuecomment-1515640138 Hello, @Lunderberg @tqchen @zhiics . Could you help check this crash? is it a bug of TVM? Many Thanks! -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #14536: [QNN] Implement 'qnn.softmax'

AndrewZhaoLuo commented on PR #14536: URL: https://github.com/apache/tvm/pull/14536#issuecomment-1515632315 Yeah it is challenging problem. I would not necessarily constrain yourself to making a single implementation that can handle all cases however. Feel free to plan for several fu

[GitHub] [tvm] maekawatoshiki closed pull request #14536: [QNN] Implement 'qnn.softmax'

maekawatoshiki closed pull request #14536: [QNN] Implement 'qnn.softmax' URL: https://github.com/apache/tvm/pull/14536 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubs

[GitHub] [tvm] maekawatoshiki commented on pull request #14536: [QNN] Implement 'qnn.softmax'

maekawatoshiki commented on PR #14536: URL: https://github.com/apache/tvm/pull/14536#issuecomment-1515628400 I took a look at the latest tensorflow implementation and noticed it is, for some reason, using float arithmetic: https://github.com/tensorflow/tensorflow/blob/98a187a8bfcdcf0c55c16f

[GitHub] [tvm] tvm-bot commented on pull request #14677: [Community] Qiang Zhang -> Reviewer

tvm-bot commented on PR #14677: URL: https://github.com/apache/tvm/pull/14677#issuecomment-1515619898 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] junrushao opened a new pull request, #14677: [Community] Qiang Zhang -> Reviewer

junrushao opened a new pull request, #14677: URL: https://github.com/apache/tvm/pull/14677 Please join me to welcome Qiang Zhang (@Johnson9009) as a new Reviewer in TVM. Qiang has been contributing extensively in Relay, runtime, CI and tests. - [Commits History](https://github.com/apa

[GitHub] [tvm] tvm-bot commented on pull request #14676: [Community] Jiajun Jiang -> Reviewer

tvm-bot commented on PR #14676: URL: https://github.com/apache/tvm/pull/14676#issuecomment-1515618000 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] junrushao opened a new pull request, #14676: [Community] Jiajun Jiang -> Reviewer

junrushao opened a new pull request, #14676: URL: https://github.com/apache/tvm/pull/14676 Please join me to welcome Jiajun Jiang (@jiangjiajun) as a new Reviewer in TVM. Jiajun has been actively maintaining the Paddlepaddle frontend in TVM. - [Commits History](https://github.com/apac

[GitHub] [tvm] maekawatoshiki commented on a diff in pull request #14536: [QNN] Implement 'qnn.softmax'

maekawatoshiki commented on code in PR #14536: URL: https://github.com/apache/tvm/pull/14536#discussion_r1172013000 ## src/relay/qnn/op/softmax.cc: ## @@ -0,0 +1,144 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements.

[tvm] branch main updated (34342ba20f -> 7c97c4eb8d)

This is an automated email from the ASF dual-hosted git repository. junrushao pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from 34342ba20f [METAL] Fix flaky memory issue due to racing (#14671) add 7c97c4eb8d [Runtime] Fix Can't "query_imports"

[GitHub] [tvm] junrushao merged pull request #14656: [Runtime] Fix Can't "query_imports" Bug of VM Executable

junrushao merged PR #14656: URL: https://github.com/apache/tvm/pull/14656 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache

[tvm] branch unity updated: [Unity] Improve WebGPU codegen for large grid (#14674)

This is an automated email from the ASF dual-hosted git repository. tqchen pushed a commit to branch unity in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/unity by this push: new 814afe6b78 [Unity] Improve WebGPU codegen for large g

[GitHub] [tvm] tqchen merged pull request #14674: [Unity] Improve WebGPU codegen for large grid

tqchen merged PR #14674: URL: https://github.com/apache/tvm/pull/14674 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or

[GitHub] [tvm] tqchen commented on pull request #14675: [Unity] Use custom hash in `BlockBuilder` to avoid hashing large constants

tqchen commented on PR #14675: URL: https://github.com/apache/tvm/pull/14675#issuecomment-1515594457 @masahi can we try keep structural equality, and cache the structural hash(can ignore bind values)(and use that to check inequality? before resorting to structural equal). I feel that should

[GitHub] [tvm] tvm-bot commented on pull request #14675: [Unity] Use custom hash in `BlockBuilder` to avoid hashing large constants

tvm-bot commented on PR #14675: URL: https://github.com/apache/tvm/pull/14675#issuecomment-1515549241 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] masahi opened a new pull request, #14675: [Unity] Use custom hash in `BlockBuilder` to avoid hashing large constants

masahi opened a new pull request, #14675: URL: https://github.com/apache/tvm/pull/14675 I found that if I use `BindParams` early in my BYOC workflow, some Relax passes become extremely slow. For example, `partition_for_cutlass` takes more than 40 sec on SD UNet if parameters are bound, whil

[GitHub] [tvm] tqchen commented on a diff in pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

tqchen commented on code in PR #14665: URL: https://github.com/apache/tvm/pull/14665#discussion_r1171892103 ## include/tvm/ir/source_map.h: ## @@ -127,6 +127,31 @@ class Span : public ObjectRef { TVM_DEFINE_OBJECT_REF_METHODS(Span, ObjectRef, SpanNode); }; +class MultiSpan

[GitHub] [tvm] areusch commented on a diff in pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

areusch commented on code in PR #14665: URL: https://github.com/apache/tvm/pull/14665#discussion_r1171882419 ## include/tvm/ir/source_map.h: ## @@ -127,6 +127,31 @@ class Span : public ObjectRef { TVM_DEFINE_OBJECT_REF_METHODS(Span, ObjectRef, SpanNode); }; +class MultiSpa

[tvm] branch unity updated: [Unity] Update docs for operators (#14659)

This is an automated email from the ASF dual-hosted git repository. tqchen pushed a commit to branch unity in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/unity by this push: new ec89242fbd [Unity] Update docs for operators (#14659)

[GitHub] [tvm] tqchen merged pull request #14659: [Unity] Update docs for operators

tqchen merged PR #14659: URL: https://github.com/apache/tvm/pull/14659 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or

[GitHub] [tvm] MasterJH5574 merged pull request #14671: [METAL] Fix flaky memory issue due to racing

MasterJH5574 merged PR #14671: URL: https://github.com/apache/tvm/pull/14671 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apa

[tvm] branch main updated (b2a7bb9ee4 -> 34342ba20f)

This is an automated email from the ASF dual-hosted git repository. ruihangl pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from b2a7bb9ee4 [MetaSchedule] Handle output cases for InlineConstantScalars (#14654) add 34342ba20f [METAL] Fix flaky m

[GitHub] [tvm] MasterJH5574 commented on pull request #14671: [METAL] Fix flaky memory issue due to racing

MasterJH5574 commented on PR #14671: URL: https://github.com/apache/tvm/pull/14671#issuecomment-1515380351 The CI issue seems to be irrelevant. So will go with merge. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the

[GitHub] [tvm] tqchen opened a new pull request, #14674: [Unity][WEBGPU] Improve WebGPU codegen for large grid

tqchen opened a new pull request, #14674: URL: https://github.com/apache/tvm/pull/14674 This PR improves webgpu codegen to handle large launch grid. Background: webgpu do not allow grid size bigger than 65535 so we have to factorize the gridDim.x when it is too big and spread it acros

[GitHub] [tvm] tvm-bot commented on pull request #14674: [Unity][WEBGPU] Improve WebGPU codegen for large grid

tvm-bot commented on PR #14674: URL: https://github.com/apache/tvm/pull/14674#issuecomment-1515376305 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[tvm-rfcs] branch main updated: [RFC] SparseTIR Dialect (#100)

This is an automated email from the ASF dual-hosted git repository. junrushao pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/tvm-rfcs.git The following commit(s) were added to refs/heads/main by this push: new 87fe5f0 [RFC] SparseTIR Dialect (#100) 87fe5f0

[GitHub] [tvm-rfcs] junrushao merged pull request #100: [RFC] SparseTIR Dialect

junrushao merged PR #100: URL: https://github.com/apache/tvm-rfcs/pull/100 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apach

[GitHub] [tvm] Lunderberg commented on pull request #14492: [TIR] Flatten SeqStmt on construction

Lunderberg commented on PR #14492: URL: https://github.com/apache/tvm/pull/14492#issuecomment-1515334121 @tvm-bot rerun -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To u

[GitHub] [tvm] masahi commented on a diff in pull request #14673: [WIP][MetaSchedule] Introduce MMA Tensor Core Multilevel Tiling

masahi commented on code in PR #14673: URL: https://github.com/apache/tvm/pull/14673#discussion_r1171759762 ## python/tvm/tir/tensor_intrin/cuda.py: ## @@ -1182,3 +1182,263 @@ def get_wmma_intrin_group( "compute": compute_intrin, "store": store_intrin, } +

[GitHub] [tvm] vinx13 merged pull request #14657: [Unity][CUTLASS] Require the residual input to have the same shape as input

vinx13 merged PR #14657: URL: https://github.com/apache/tvm/pull/14657 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or

[tvm] branch unity updated: [Unity][CUTLASS] Require the residual input to have the same shape as input (#14657)

This is an automated email from the ASF dual-hosted git repository. wuwei pushed a commit to branch unity in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/unity by this push: new f7835a6f80 [Unity][CUTLASS] Require the residual input

[tvm] branch main updated: [MetaSchedule] Handle output cases for InlineConstantScalars (#14654)

This is an automated email from the ASF dual-hosted git repository. wuwei pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/main by this push: new b2a7bb9ee4 [MetaSchedule] Handle output cases for Inlin

[GitHub] [tvm] vinx13 merged pull request #14654: [MetaSchedule] Handle output cases for InlineConstantScalars

vinx13 merged PR #14654: URL: https://github.com/apache/tvm/pull/14654 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or

[tvm] branch unity updated: [Unity] Add rewriting for CUDA graph capturing (#14513)

This is an automated email from the ASF dual-hosted git repository. wuwei pushed a commit to branch unity in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/unity by this push: new bb1e10c81f [Unity] Add rewriting for CUDA graph captur

[GitHub] [tvm] vinx13 merged pull request #14513: [Unity] Add rewriting for CUDA graph capturing

vinx13 merged PR #14513: URL: https://github.com/apache/tvm/pull/14513 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.or

[GitHub] [tvm] tvm-bot commented on pull request #14673: [WIP][MetaSchedule] Introduce MMA Tensor Core Multilevel Tiling

tvm-bot commented on PR #14673: URL: https://github.com/apache/tvm/pull/14673#issuecomment-1515217433 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] cblmemo opened a new pull request, #14673: [WIP][MetaSchedule] Introduce MMA Tensor Core Multilevel Tiling

cblmemo opened a new pull request, #14673: URL: https://github.com/apache/tvm/pull/14673 This PR introduces mma in the multilevel tiling tensor core. This PR is still a WIP. TODO: add the permuted layout. For the current benchmark result without permuted layout, please re

[GitHub] [tvm] tvm-bot commented on pull request #14672: [ci] Add sccache logging

tvm-bot commented on PR #14672: URL: https://github.com/apache/tvm/pull/14672#issuecomment-1515076681 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] tqchen commented on a diff in pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

tqchen commented on code in PR #14665: URL: https://github.com/apache/tvm/pull/14665#discussion_r1171622886 ## include/tvm/ir/source_map.h: ## @@ -127,6 +127,31 @@ class Span : public ObjectRef { TVM_DEFINE_OBJECT_REF_METHODS(Span, ObjectRef, SpanNode); }; +class MultiSpan

[GitHub] [tvm] areusch commented on a diff in pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

areusch commented on code in PR #14665: URL: https://github.com/apache/tvm/pull/14665#discussion_r1171549494 ## src/relax/transform/fuse_ops.cc: ## @@ -730,7 +751,8 @@ class OperatorFusor : public ExprMutator { // - If this binding is an output binding, emit an output va

[GitHub] [tvm] areusch commented on pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

areusch commented on PR #14665: URL: https://github.com/apache/tvm/pull/14665#issuecomment-1514978253 @Hzfengsy yes it should work for FuseOpsByPattern (I've updated CompositeFunctionAnnotator to make this work). @junrushao I agree there are limitations here when it comes to Spans in

[GitHub] [tvm] Ubospica commented on a diff in pull request #14670: [Unity][Training] Optimizer library

Ubospica commented on code in PR #14670: URL: https://github.com/apache/tvm/pull/14670#discussion_r1171546946 ## python/tvm/relax/__init__.py: ## @@ -88,8 +88,10 @@ from . import op from . import struct_info from . import backend +from . import training from . import fronten

[GitHub] [tvm] areusch commented on a diff in pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

areusch commented on code in PR #14665: URL: https://github.com/apache/tvm/pull/14665#discussion_r1171539372 ## include/tvm/ir/source_map.h: ## @@ -127,6 +127,31 @@ class Span : public ObjectRef { TVM_DEFINE_OBJECT_REF_METHODS(Span, ObjectRef, SpanNode); }; +class MultiSpa

[GitHub] [tvm] FranklandJack commented on a diff in pull request #14483: [TOPI][Target] Add SVE specific convolution

FranklandJack commented on code in PR #14483: URL: https://github.com/apache/tvm/pull/14483#discussion_r1171465090 ## tests/cpp/target/parsers/aprofile_test.cc: ## @@ -290,9 +290,28 @@ TEST(AProfileParser, ArchVersionInvalidLetter) { ASSERT_EQ(Downcast(features.at("has_dotpro

[GitHub] [tvm] tvm-bot commented on pull request #14671: [METAL] Fix flaky memory issue due to racing

tvm-bot commented on PR #14671: URL: https://github.com/apache/tvm/pull/14671#issuecomment-1514871031 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] tqchen opened a new pull request, #14671: [METAL] Fix flaky memory issue due to racing

tqchen opened a new pull request, #14671: URL: https://github.com/apache/tvm/pull/14671 This PR aims to fix a relatively rare flaky race issuse when we are attempting to purge memory and the buffer is still in use in command buffer -- This is an automated message from the Apache Git Servi

[GitHub] [tvm] SiriusNEO commented on a diff in pull request #14670: [Unity][Training] Optimizer library

SiriusNEO commented on code in PR #14670: URL: https://github.com/apache/tvm/pull/14670#discussion_r1171448183 ## python/tvm/relax/__init__.py: ## @@ -88,8 +88,10 @@ from . import op from . import struct_info from . import backend +from . import training from . import fronte

[GitHub] [tvm] Mousius commented on a diff in pull request #14483: [TOPI][Target] Add SVE specific convolution

Mousius commented on code in PR #14483: URL: https://github.com/apache/tvm/pull/14483#discussion_r1171444769 ## tests/cpp/target/parsers/aprofile_test.cc: ## @@ -290,9 +290,28 @@ TEST(AProfileParser, ArchVersionInvalidLetter) { ASSERT_EQ(Downcast(features.at("has_dotprod")),

[GitHub] [tvm] FranklandJack commented on a diff in pull request #14519: [TOPI] Vectorize depthwise conv2d output operator

FranklandJack commented on code in PR #14519: URL: https://github.com/apache/tvm/pull/14519#discussion_r1171436553 ## python/tvm/topi/arm_cpu/depthwise_conv2d.py: ## @@ -394,7 +394,8 @@ def schedule_conv_out(out): ci_outer, ci_inner = s[out].split(ci, 4)

[GitHub] [tvm] tqchen commented on a diff in pull request #14665: [Unity] Add MultiSpan and attach it to CallNode produced by FuseOps

tqchen commented on code in PR #14665: URL: https://github.com/apache/tvm/pull/14665#discussion_r1171376249 ## include/tvm/ir/source_map.h: ## @@ -127,6 +127,31 @@ class Span : public ObjectRef { TVM_DEFINE_OBJECT_REF_METHODS(Span, ObjectRef, SpanNode); }; +class MultiSpan

[GitHub] [tvm] FranklandJack commented on a diff in pull request #14483: [TOPI][Target] Add SVE specific convolution

FranklandJack commented on code in PR #14483: URL: https://github.com/apache/tvm/pull/14483#discussion_r1171333756 ## tests/cpp/target/parsers/aprofile_test.cc: ## @@ -290,9 +290,28 @@ TEST(AProfileParser, ArchVersionInvalidLetter) { ASSERT_EQ(Downcast(features.at("has_dotpro

[GitHub] [tvm] dmilosevic252 commented on pull request #14580: [Relay] Simplify the square of a binomial

dmilosevic252 commented on PR #14580: URL: https://github.com/apache/tvm/pull/14580#issuecomment-1514721396 @tvm-bot rerun -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. T

[GitHub] [tvm] Lunderberg commented on pull request #14492: [TIR] Flatten SeqStmt on construction

Lunderberg commented on PR #14492: URL: https://github.com/apache/tvm/pull/14492#issuecomment-1514709413 No worries (this has been on my back burner as well). Main has been merged into the PR branch to resolve conflicts, and now running through CI. -- This is an automated message from th

[GitHub] [tvm] Aleksei-grovety commented on pull request #14629: [microNPU][ETHOSU] Softmax int8 legalization support

Aleksei-grovety commented on PR #14629: URL: https://github.com/apache/tvm/pull/14629#issuecomment-1514708044 There is a problem with run microTVM demos, which consists in a greatly increased time to execute lower_to_te, I'm trying to find out the reason for this -- This is an automated

[GitHub] [tvm] Mousius commented on a diff in pull request #14483: [TOPI][Target] Add SVE specific convolution

Mousius commented on code in PR #14483: URL: https://github.com/apache/tvm/pull/14483#discussion_r1171312805 ## tests/cpp/target/parsers/aprofile_test.cc: ## @@ -290,9 +290,28 @@ TEST(AProfileParser, ArchVersionInvalidLetter) { ASSERT_EQ(Downcast(features.at("has_dotprod")),

[GitHub] [tvm] Aleksei-grovety commented on a diff in pull request #14536: [QNN] Implement 'qnn.softmax'

Aleksei-grovety commented on code in PR #14536: URL: https://github.com/apache/tvm/pull/14536#discussion_r1171182996 ## src/relay/qnn/op/softmax.cc: ## @@ -0,0 +1,144 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements

[GitHub] [tvm] tvm-bot commented on pull request #14670: [Unity][Training] Optimizer library

tvm-bot commented on PR #14670: URL: https://github.com/apache/tvm/pull/14670#issuecomment-1514521919 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] Ubospica opened a new pull request, #14670: [Unity][Training] Optimizer library

Ubospica opened a new pull request, #14670: URL: https://github.com/apache/tvm/pull/14670 This PR adds the optimizer library to support the Relax training workflow. Optimizer library generates relax functions representing common optimizers, such as SGD and Adam. Taking SGD as a

[GitHub] [tvm] tvm-bot commented on pull request #14669: [Unity] Fix `DataflowReshapeRewrite` when input has multiple buffers from tuple

tvm-bot commented on PR #14669: URL: https://github.com/apache/tvm/pull/14669#issuecomment-1514514820 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] masahi opened a new pull request, #14669: [Unity] Fix `DataflowReshapeRewrite` when input has multiple buffers from tuple

masahi opened a new pull request, #14669: URL: https://github.com/apache/tvm/pull/14669 Currently, `DataflowReshapeRewrite` expects that `call_tir(reshape, (...))` gets only one input: https://github.com/apache/tvm/blob/unity/src/relax/transform/rewrite_dataflow_reshape.cc#L66-L67

[GitHub] [tvm] tvm-bot commented on pull request #14668: [Unity][Training] Loss functions and AppendLoss pass

tvm-bot commented on PR #14668: URL: https://github.com/apache/tvm/pull/14668#issuecomment-1514504975 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] Ubospica opened a new pull request, #14668: [Unity][Training] Loss functions and AppendLoss pass

Ubospica opened a new pull request, #14668: URL: https://github.com/apache/tvm/pull/14668 This PR adds two components to support the Relax training workflow: loss function library and AppendLoss pass. - Loss function library. A library that generates relax functions representing common l

[GitHub] [tvm] FranklandJack commented on pull request #14519: [TOPI] Vectorize depthwise conv2d output operator

FranklandJack commented on PR #14519: URL: https://github.com/apache/tvm/pull/14519#issuecomment-1514455412 > > I think at present though we don't do any testing for Arm Cortex A targets (see [here](https://github.com/apache/tvm/blob/main/tests/python/relay/strategy/arm_cpu/test_generalized

[GitHub] [tvm] leandron closed pull request #11249: [TFLite] Support quantised SQUARED_DIFFERENCE operator

leandron closed pull request #11249: [TFLite] Support quantised SQUARED_DIFFERENCE operator URL: https://github.com/apache/tvm/pull/11249 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific

[GitHub] [tvm] leandron commented on pull request #11249: [TFLite] Support quantised SQUARED_DIFFERENCE operator

leandron commented on PR #11249: URL: https://github.com/apache/tvm/pull/11249#issuecomment-1514421810 Closing in favour of https://github.com/apache/tvm/pull/14667 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the U

[GitHub] [tvm] ysh329 commented on issue #14645: [Release] v0.12.0 Release Candidate Notes

ysh329 commented on issue #14645: URL: https://github.com/apache/tvm/issues/14645#issuecomment-1514382528 > Very minor, there is a typo "Reviwer" -> Reviewer. Again, **thanks** for putting the work, I created this changelog once and it is not a simple job. Fixed. Thanks a lot! 👍 --

[GitHub] [tvm] tvm-bot commented on pull request #14667: [TFLite][Frontend] Support for quantized squared difference

tvm-bot commented on PR #14667: URL: https://github.com/apache/tvm/pull/14667#issuecomment-1514360507 Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from

[GitHub] [tvm] ashutosh-arm opened a new pull request, #14667: [TFLite][Frontend] Support for quantized squared difference

ashutosh-arm opened a new pull request, #14667: URL: https://github.com/apache/tvm/pull/14667 In line with the TFLu implementation, this commit enables mapping of quantized squared difference to Relay's float32 squred difference. cc @leandron -- This is an automated message f

[GitHub] [tvm] leandron commented on issue #14645: [Release] v0.12.0 Release Candidate Notes

leandron commented on issue #14645: URL: https://github.com/apache/tvm/issues/14645#issuecomment-1514302097 Very minor, there is a typo "Reviwer" -> Reviewer. Again, **thanks** for putting the work, I created this changelog once and it is not a simple job. -- This is an automated message

[tvm] branch main updated (8e4cedc4e3 -> fc7ca1f503)

This is an automated email from the ASF dual-hosted git repository. driazati pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git from 8e4cedc4e3 [Docker] Support rootless docker when using docker/bash.sh (#14590) add fc7ca1f503 [tests][scripts][rele

[GitHub] [tvm] driazati merged pull request #14653: [tests][scripts][release] Optimize release note script about categories etc

driazati merged PR #14653: URL: https://github.com/apache/tvm/pull/14653 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.

[GitHub] [tvm] masahi closed issue #14663: [Tracking Issue]

masahi closed issue #14663: [Tracking Issue] URL: https://github.com/apache/tvm/issues/14663 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-uns