[GitHub] [tvm] MasterJH5574 commented on a change in pull request #10081: [MetaSchedule][M4a] Rewrite-Cooperative-Fetch

MasterJH5574 commented on a change in pull request #10081: URL: https://github.com/apache/tvm/pull/10081#discussion_r795006807 ## File path: src/meta_schedule/postproc/rewrite_cooperative_fetch.cc ## @@ -0,0 +1,158 @@ +/* + * Licensed to the Apache Software Foundation (ASF) und

[GitHub] [tvm] Hzfengsy commented on a change in pull request #10081: [MetaSchedule][M4a] Rewrite-Cooperative-Fetch

Hzfengsy commented on a change in pull request #10081: URL: https://github.com/apache/tvm/pull/10081#discussion_r795006308 ## File path: src/meta_schedule/postproc/rewrite_cooperative_fetch.cc ## @@ -0,0 +1,158 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under o

[GitHub] [tvm] Hzfengsy commented on a change in pull request #10081: [MetaSchedule][M4a] Rewrite-Cooperative-Fetch

Hzfengsy commented on a change in pull request #10081: URL: https://github.com/apache/tvm/pull/10081#discussion_r795006297 ## File path: src/meta_schedule/postproc/rewrite_cooperative_fetch.cc ## @@ -0,0 +1,158 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under o

[tvm] branch main updated (4d0dac3 -> 21154c2)

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 4d0dac3 [MetaSchedule][M4a] Mutator: Mutate-Tile-Size (#10092) add 21154c2 [TE] Fix Const Int bound analysis to

[GitHub] [tvm] junrushao1994 merged pull request #10102: [TE] Fix Const Int bound analysis to handle uints for division

junrushao1994 merged pull request #10102: URL: https://github.com/apache/tvm/pull/10102 -- 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-un

[GitHub] [tvm-rfcs] masahi merged pull request #52: [RFC][OpenCLML] OpenCLML integration as BYOC

masahi merged pull request #52: URL: https://github.com/apache/tvm-rfcs/pull/52 -- 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-rfcs] branch main updated: [RFC][OpenCLML] OpenCLML integration as BYOC (#52)

This is an automated email from the ASF dual-hosted git repository. masahi 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 f5ef65f [RFC][OpenCLML] OpenCLML integration as B

[tvm] branch main updated: [MetaSchedule][M4a] Mutator: Mutate-Tile-Size (#10092)

This is an automated email from the ASF dual-hosted git repository. syfeng 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 4d0dac3 [MetaSchedule][M4a] Mutator: Mutate-Tile-Size

[GitHub] [tvm] Hzfengsy merged pull request #10092: [MetaSchedule][M4a] Mutator: Mutate-Tile-Size

Hzfengsy merged pull request #10092: URL: https://github.com/apache/tvm/pull/10092 -- 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-unsubsc

[tvm] branch ci-docker-staging updated: remove gcn tutorial workaround

This is an automated email from the ASF dual-hosted git repository. masahi pushed a commit to branch ci-docker-staging in repository https://gitbox.apache.org/repos/asf/tvm.git The following commit(s) were added to refs/heads/ci-docker-staging by this push: new 8ddf403 remove gcn tutorial

[tvm] branch ci-docker-staging updated (6c52594 -> 42e1dc3)

This is an automated email from the ASF dual-hosted git repository. masahi pushed a change to branch ci-docker-staging in repository https://gitbox.apache.org/repos/asf/tvm.git. discard 6c52594 validating ci_gpu:20220128-070420-fa317edf7 new 42e1dc3 validating ci_gpu:20220128-070420

[tvm] 01/01: validating ci_gpu:20220128-070420-fa317edf7

validating ci_gpu:20220128-070420-fa317edf7 --- Jenkinsfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Jenkinsfile b/Jenkinsfile index 1cd8575..9664114 100755 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -46,7 +46,7 @@ import org.jenkinsci.plugins.pipeline.modeldefinition.Utils

[tvm] branch ci-docker-staging updated (07177cc -> 6c52594)

] Pipeline Executor Sequential execution (#10082) add 85d42f8 [MetaSchedule][M4a] Mutator: Mutate Parallel (#10096) add 6a274af [Hexagon] Update hexagon API build instruction and cleanup hexagon_proxy_rpc (#10068) new 6c52594 validating ci_gpu:20220128-070420-fa317edf7 This update

[tvm] 01/01: validating ci_gpu:20220128-070420-fa317edf7

validating ci_gpu:20220128-070420-fa317edf7 --- Jenkinsfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Jenkinsfile b/Jenkinsfile index 1cd8575..55c90c2 100755 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -46,7 +46,7 @@ import org.jenkinsci.plugins.pipeline.modeldefinition.Utils

[GitHub] [tvm] huajsj closed pull request #9494: [WIP][Runtime] Pipeline Executor Add Set and Get Input/Output interfaces.

huajsj closed pull request #9494: URL: https://github.com/apache/tvm/pull/9494 -- 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...

[GitHub] [tvm] vinx13 commented on issue #10103: [Bug] tests/python/contrib/test_popen_pool.py::test_popen_pool_executor_async is flaky

vinx13 commented on issue #10103: URL: https://github.com/apache/tvm/issues/10103#issuecomment-1024766357 Which one finishing first doesn't really matter, it is merely to show that they are concurrent. I think it is fine to remove this test case. @shingjan would you like to send a PR? --

[GitHub] [tvm] shingjan commented on issue #10103: [Bug] tests/python/contrib/test_popen_pool.py::test_popen_pool_executor_async is flaky

shingjan commented on issue #10103: URL: https://github.com/apache/tvm/issues/10103#issuecomment-1024765277 the `timeout` parameter could be the source of flakiness. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use th

[GitHub] [tvm] junrushao1994 commented on a change in pull request #10055: [MetaSchedule][M4b] Testcases for TensorRT builder/runner

junrushao1994 commented on a change in pull request #10055: URL: https://github.com/apache/tvm/pull/10055#discussion_r794951021 ## File path: tests/python/unittest/test_meta_schedule_byoc_tensorrt.py ## @@ -0,0 +1,251 @@ +# Licensed to the Apache Software Foundation (ASF) under

[GitHub] [tvm] driazati opened a new issue #10103: [Bug] tests/python/contrib/test_popen_pool.py::test_popen_pool_executor_async is flaky

driazati opened a new issue #10103: URL: https://github.com/apache/tvm/issues/10103 This test (from #8851) seems flaky by design, does it really matter which task finishes first in absolute time? example failure on main: https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail

[GitHub] [tvm] anwang2009 commented on a change in pull request #10089: [Relay][Pass] Add a relay pass to extract fake quantized ops

anwang2009 commented on a change in pull request #10089: URL: https://github.com/apache/tvm/pull/10089#discussion_r794938851 ## File path: src/relay/analysis/extract_fake_quantized_ops.cc ## @@ -0,0 +1,108 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one +

[tvm] branch main updated (85d42f8 -> 6a274af)

This is an automated email from the ASF dual-hosted git repository. lunderberg pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 85d42f8 [MetaSchedule][M4a] Mutator: Mutate Parallel (#10096) add 6a274af [Hexagon] Update hexagon API build in

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #10098: [Op][Topi] 5 ops can accept unsigned integers as indices

AndrewZhaoLuo commented on pull request #10098: URL: https://github.com/apache/tvm/pull/10098#issuecomment-1024725194 https://github.com/apache/tvm/pull/10102 <-- fix here. There is another error now with unravel index now though. You can try digging more, or you can drop unravel_index for

[GitHub] [tvm] Lunderberg merged pull request #10068: [Hexagon] Update hexagon API build instruction and cleanup hexagon_proxy_rpc

Lunderberg merged pull request #10068: URL: https://github.com/apache/tvm/pull/10068 -- 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-unsub

[GitHub] [tvm] AndrewZhaoLuo opened a new pull request #10102: [TE] Fix Const Int bound analysis to handle uints for division

AndrewZhaoLuo opened a new pull request #10102: URL: https://github.com/apache/tvm/pull/10102 The analyzer tries to see the range an integer variable can represent. Specifically, if the range of a divisor contained zero it would just sort of be ignored (assuming probably that run time chec

[GitHub] [tvm] AndrewZhaoLuo edited a comment on pull request #10098: [Op][Topi] 5 ops can accept unsigned integers as indices

AndrewZhaoLuo edited a comment on pull request #10098: URL: https://github.com/apache/tvm/pull/10098#issuecomment-1024670599 Debugging I've worked the issue to likely `src/tir/transforms/narrow_datatype.cc`, my guess analysis here does not take into account uint types -- This is an auto

[GitHub] [tvm] AndrewZhaoLuo commented on pull request #10098: [Op][Topi] 5 ops can accept unsigned integers as indices

AndrewZhaoLuo commented on pull request #10098: URL: https://github.com/apache/tvm/pull/10098#issuecomment-1024670599 Debugging I've worked the issue to likely `src/tir/transforms/narrow_datatype.cc` -- This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [tvm] mehrdadh commented on pull request #9631: [Hexagon] Add RPC Mechanism for Hexagon

mehrdadh commented on pull request #9631: URL: https://github.com/apache/tvm/pull/9631#issuecomment-1024660830 @kparzysz-quic thanks for pointing out these issues. I will address these in the next PR for AOT executor. -- This is an automated message from the Apache Git Service. To respon

[GitHub] [tvm] Raghav-Chakravarthy opened a new pull request #10101: [Code Style Change] - Changed Vulkan code to match TVM conventions.

Raghav-Chakravarthy opened a new pull request #10101: URL: https://github.com/apache/tvm/pull/10101 Code Style Change - Changed Vulkan code to match TVM conventions. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use

[GitHub] [tvm] junrushao1994 commented on pull request #10099: [TVMScript] Support T.buffer_decl using data pointer from Let/Allocate

junrushao1994 commented on pull request #10099: URL: https://github.com/apache/tvm/pull/10099#issuecomment-1024638501 CC: @Hzfengsy @vinx13 @spectrometerHBH -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL abo

[tvm] branch main updated (80d4d05 -> 85d42f8)

This is an automated email from the ASF dual-hosted git repository. bohan pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 80d4d05 [Runtime][PipelineExecutor] Pipeline Executor Sequential execution (#10082) add 85d42f8 [MetaSchedule][M4a]

[GitHub] [tvm] spectrometerHBH merged pull request #10096: [MetaSchedule][M4a] Mutator: Mutate Parallel

spectrometerHBH merged pull request #10096: URL: https://github.com/apache/tvm/pull/10096 -- 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-

[tvm] branch main updated (3af9c30 -> 80d4d05)

This is an automated email from the ASF dual-hosted git repository. comaniac pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 3af9c30 [microTVM] Update Zephyr to 2.7 (#10094) add 80d4d05 [Runtime][PipelineExecutor] Pipeline Executor Sequen

[GitHub] [tvm] comaniac merged pull request #10082: [Runtime][PipelineExecutor] Pipeline Executor Sequential execution

comaniac merged pull request #10082: URL: https://github.com/apache/tvm/pull/10082 -- 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-unsubsc

[GitHub] [tvm] comaniac commented on pull request #10082: [Runtime][PipelineExecutor] Pipeline Executor Sequential execution

comaniac commented on pull request #10082: URL: https://github.com/apache/tvm/pull/10082#issuecomment-1024617361 Thanks @huajsj -- 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 comm

[GitHub] [tvm] AndrewZhaoLuo commented on a change in pull request #10089: [Relay][Pass] Add a relay pass to extract fake quantized ops

AndrewZhaoLuo commented on a change in pull request #10089: URL: https://github.com/apache/tvm/pull/10089#discussion_r794764050 ## File path: tests/python/relay/test_analysis_extract_fake_quantized_ops.py ## @@ -0,0 +1,142 @@ +# Licensed to the Apache Software Foundation (ASF)

[GitHub] [tvm] huajsj commented on pull request #10082: [Runtime][PipelineExecutor] Pipeline Executor Sequential execution

huajsj commented on pull request #10082: URL: https://github.com/apache/tvm/pull/10082#issuecomment-1024582700 @comaniac , thanks for the fix, all review comments addressed, please take a look. -- This is an automated message from the Apache Git Service. To respond to the message, please

[GitHub] [tvm] grant-arm commented on a change in pull request #10095: [microTVM] Include standalone_crt dependencies in MLF

grant-arm commented on a change in pull request #10095: URL: https://github.com/apache/tvm/pull/10095#discussion_r794829776 ## File path: python/tvm/micro/model_library_format.py ## @@ -317,6 +321,16 @@ def _export_graph_model_library_format( "style": "full-model",

[GitHub] [tvm] ashutosh-arm opened a new pull request #10100: [CMSIS-NN] Convert scalar constants to tensor constants

ashutosh-arm opened a new pull request #10100: URL: https://github.com/apache/tvm/pull/10100 * Scalar to tensor constant pass -Pass that converts scalar constants into tensor constants for CMSIS-NN binary ops -Updated extract_constants pass to support constants at any position in

[GitHub] [tvm] d-smirnov commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

d-smirnov commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r794780497 ## File path: include/tvm/ir/module.h ## @@ -349,6 +390,9 @@ class IRModuleNode : public Object { */ std::unordered_set import_set_; friend class

[GitHub] [tvm] d-smirnov commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

d-smirnov commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r794775246 ## File path: src/ir/module.cc ## @@ -219,6 +224,64 @@ void IRModuleNode::AddUnchecked(const GlobalVar& var, const BaseFunc& func) { global_var_map_.Set

[GitHub] [tvm] d-smirnov commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

d-smirnov commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r794771519 ## File path: src/relay/transforms/fuse_ops.cc ## @@ -1017,18 +1050,30 @@ class FuseMutator : private MixedModeMutator { } }; -Expr FuseOps(const Expr

[GitHub] [tvm] d-smirnov commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

d-smirnov commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r794771308 ## File path: src/relay/transforms/fuse_ops.cc ## @@ -809,8 +811,31 @@ std::vector GraphPartitioner::Partition( class FuseMutator : private MixedModeMut

[GitHub] [tvm] tqchen commented on pull request #9980: [VirtualMachine] fix raw pointer using by VirtualMachine

tqchen commented on pull request #9980: URL: https://github.com/apache/tvm/pull/9980#issuecomment-1024509306 @vvchernov would be great if you can confirm my latest comment, which should be a better fix ``` Sorry for the delayed reply. sptr_to_self actually is guanranteed to point

[GitHub] [tvm] Lunderberg commented on pull request #10099: [TVMScript] Support T.buffer_decl using data pointer from Let/Allocate

Lunderberg commented on pull request #10099: URL: https://github.com/apache/tvm/pull/10099#issuecomment-1024496920 Updated to resolve lint errors. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to

[GitHub] [tvm] manupa-arm edited a comment on pull request #10022: [microNPU] enable USMP

manupa-arm edited a comment on pull request #10022: URL: https://github.com/apache/tvm/pull/10022#issuecomment-1018780014 ~~This is blocked on #9929 and #9951~~ -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL

[GitHub] [tvm] leandron commented on a change in pull request #10095: [microTVM] Include standalone_crt dependencies in MLF

leandron commented on a change in pull request #10095: URL: https://github.com/apache/tvm/pull/10095#discussion_r794740072 ## File path: python/tvm/micro/model_library_format.py ## @@ -287,6 +288,9 @@ def reset(tarinfo): return tarinfo tar_f.add(str(sour

[GitHub] [tvm] Lunderberg opened a new pull request #10099: [TVMScript] Support T.buffer_decl using data pointer from Let/Allocate

Lunderberg opened a new pull request #10099: URL: https://github.com/apache/tvm/pull/10099 Supporting this use case required additional support on both the parsing side and the tvmscript printing side. * Parse pointer objects `T.Ptr[T.int32]` in TVMScript. Previously, there was no

[GitHub] [tvm] MasterJH5574 commented on a change in pull request #10081: [MetaSchedule][M4a] Rewrite-Cooperative-Fetch

MasterJH5574 commented on a change in pull request #10081: URL: https://github.com/apache/tvm/pull/10081#discussion_r794733853 ## File path: src/meta_schedule/postproc/rewrite_cooperative_fetch.cc ## @@ -0,0 +1,158 @@ +/* + * Licensed to the Apache Software Foundation (ASF) und

[tvm] branch main updated (82d4d0f -> 3af9c30)

This is an automated email from the ASF dual-hosted git repository. leandron pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 82d4d0f [Relay] fix incorrect binding of Lets in ANF conversion (#10078) add 3af9c30 [microTVM] Update Zephyr to

[GitHub] [tvm] leandron commented on pull request #10094: [microTVM] Update Zephyr to 2.7

leandron commented on pull request #10094: URL: https://github.com/apache/tvm/pull/10094#issuecomment-1024446055 Thanks @Mousius! This is merged now. -- 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

[GitHub] [tvm] leandron merged pull request #10094: [microTVM] Update Zephyr to 2.7

leandron merged pull request #10094: URL: https://github.com/apache/tvm/pull/10094 -- 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-unsubsc

[tvm] branch ci-docker-staging updated (a45c68d -> 07177cc)

This is an automated email from the ASF dual-hosted git repository. mousius pushed a change to branch ci-docker-staging in repository https://gitbox.apache.org/repos/asf/tvm.git. omit a45c68d [CI] Move docs precheck to docs: GPU add f93f2a6 [microTVM][tvmc] Add TVMC Micro tutorial for

[GitHub] [tvm] Mousius closed pull request #10093: [CI] Move docs precheck to docs: GPU

Mousius closed pull request #10093: URL: https://github.com/apache/tvm/pull/10093 -- 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

[GitHub] [tvm] Mousius commented on pull request #10093: [CI] Move docs precheck to docs: GPU

Mousius commented on pull request #10093: URL: https://github.com/apache/tvm/pull/10093#issuecomment-1024432922 Ah, didn't see #9971 -- 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

[tvm] branch main updated (7b9fd1e -> 82d4d0f)

This is an automated email from the ASF dual-hosted git repository. mbrookhart pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from 7b9fd1e [microNPU] Removing constant args from PrimFunc (#9951) add 82d4d0f [Relay] fix incorrect binding of Le

[GitHub] [tvm] mbrookhart commented on pull request #10078: [Relay] fix incorrect binding of Lets in ANF conversion

mbrookhart commented on pull request #10078: URL: https://github.com/apache/tvm/pull/10078#issuecomment-1024432339 Thanks @altanh @electriclilies @mbs-octoml -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL ab

[GitHub] [tvm] mbrookhart merged pull request #10078: [Relay] fix incorrect binding of Lets in ANF conversion

mbrookhart merged pull request #10078: URL: https://github.com/apache/tvm/pull/10078 -- 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-unsub

[GitHub] [tvm] FZYUAN-1 opened a new pull request #10098: [Op][Topi] 5 ops can accept unsigned integers as indices

FZYUAN-1 opened a new pull request #10098: URL: https://github.com/apache/tvm/pull/10098 Hello community, This PR links to #10064. 5 ops (scatter, scatter_add, scatter_nd, unravel_index, adv_index) accept unsigned integers as indices. Somehow unravel_index with uint indices may rais

[GitHub] [tvm] tmoreau89 commented on pull request #9980: [VirtualMachine] fix raw pointer using by VirtualMachine

tmoreau89 commented on pull request #9980: URL: https://github.com/apache/tvm/pull/9980#issuecomment-1024377962 At first sight the CI error seems completely unrelated to the PR... -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitH

[GitHub] [tvm-rfcs] gayatripk1 removed a comment on pull request #6: [RFC] [Relay] Automatic Mixed Precision Pass

gayatripk1 removed a comment on pull request #6: URL: https://github.com/apache/tvm-rfcs/pull/6#issuecomment-1024374936 @AndrewZhaoLuo, When we enable mixed precision arithmetic, the output types of a model depends on the last operation in the relay IR, ie, if it convolution, it will be in

[GitHub] [tvm] tmoreau89 commented on pull request #9980: [VirtualMachine] fix raw pointer using by VirtualMachine

tmoreau89 commented on pull request #9980: URL: https://github.com/apache/tvm/pull/9980#issuecomment-1024377474 @driazati does this error seem like a flaky issue to you? > worker 'gw3' crashed while running 'tests/python/contrib/test_ethosu/test_codegen.py::test_mean[ifm_shape8-axis8-Fal

[GitHub] [tvm-rfcs] gayatripk1 commented on pull request #6: [RFC] [Relay] Automatic Mixed Precision Pass

gayatripk1 commented on pull request #6: URL: https://github.com/apache/tvm-rfcs/pull/6#issuecomment-1024374936 @AndrewZhaoLuo, When we enable mixed precision arithmetic, the output types of a model depends on the last operation in the relay IR, ie, if it convolution, it will be in float16

[tvm] branch main updated (e6af874 -> 7b9fd1e)

This is an automated email from the ASF dual-hosted git repository. mousius pushed a change to branch main in repository https://gitbox.apache.org/repos/asf/tvm.git. from e6af874 [PyTorch] Fix rsub type (#10090) add 7b9fd1e [microNPU] Removing constant args from PrimFunc (#9951) No n

[GitHub] [tvm] Mousius commented on pull request #9951: [microNPU] Removing constant args from PrimFunc

Mousius commented on pull request #9951: URL: https://github.com/apache/tvm/pull/9951#issuecomment-1024310329 Thanks @manupa-arm / @mbaret - I declare this PR merged! :smile_cat: -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitH

[GitHub] [tvm] Mousius merged pull request #9951: [microNPU] Removing constant args from PrimFunc

Mousius merged pull request #9951: URL: https://github.com/apache/tvm/pull/9951 -- 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..

[GitHub] [tvm] prateek9623 opened a new pull request #10097: [CMake] add support for find_package

prateek9623 opened a new pull request #10097: URL: https://github.com/apache/tvm/pull/10097 This PR adds the following features: - add target tvm to a CMake export group, which enables its linkage with another target with another target group. - adds support for cmake find_package

[GitHub] [tvm] jinhongyii opened a new pull request #10096: [MetaSchedule][M4a] Mutator: Mutate Parallel

jinhongyii opened a new pull request #10096: URL: https://github.com/apache/tvm/pull/10096 This PR is one of the mutator for MetaSchedule. Co-authored-by: Junru Shao junrushao1...@gmail.com Co-authored-by: Xiyou Zhou xi...@octoml.ai Co-authored-by: Bohan Hou spectromet...@gmail.

[GitHub] [tvm] manupa-arm commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

manupa-arm commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r794475756 ## File path: src/relay/transforms/fuse_ops.cc ## @@ -1017,18 +1050,30 @@ class FuseMutator : private MixedModeMutator { } }; -Expr FuseOps(const Exp

[GitHub] [tvm] manupa-arm commented on a change in pull request #8509: [TIR] Tir constants integration into compilation pipeline

manupa-arm commented on a change in pull request #8509: URL: https://github.com/apache/tvm/pull/8509#discussion_r794475756 ## File path: src/relay/transforms/fuse_ops.cc ## @@ -1017,18 +1050,30 @@ class FuseMutator : private MixedModeMutator { } }; -Expr FuseOps(const Exp

[GitHub] [tvm] grant-arm commented on a change in pull request #10095: [mircoTVM] Include standalone_crt dependencies in MLF

grant-arm commented on a change in pull request #10095: URL: https://github.com/apache/tvm/pull/10095#discussion_r794448151 ## File path: apps/microtvm/ethosu/Makefile ## @@ -19,13 +19,13 @@ # Setup build environment BUILD_DIR := build -STANDALONE_CRT_PATH := $(shell python

[GitHub] [tvm-rfcs] manupa-arm commented on a change in pull request #46: Module Based Model Runtime for AOT

manupa-arm commented on a change in pull request #46: URL: https://github.com/apache/tvm-rfcs/pull/46#discussion_r794433151 ## File path: rfcs/0046-module-based-model-runtime-for-aot.md ## @@ -0,0 +1,348 @@ +# Module-based Model Runtime Interface for AOT + +- Feature Name: modu

[GitHub] [tvm-rfcs] manupa-arm commented on a change in pull request #46: Module Based Model Runtime for AOT

manupa-arm commented on a change in pull request #46: URL: https://github.com/apache/tvm-rfcs/pull/46#discussion_r794433151 ## File path: rfcs/0046-module-based-model-runtime-for-aot.md ## @@ -0,0 +1,348 @@ +# Module-based Model Runtime Interface for AOT + +- Feature Name: modu

[GitHub] [tvm-rfcs] manupa-arm commented on a change in pull request #46: Module Based Model Runtime for AOT

manupa-arm commented on a change in pull request #46: URL: https://github.com/apache/tvm-rfcs/pull/46#discussion_r794407546 ## File path: rfcs/0046-module-based-model-runtime-for-aot.md ## @@ -0,0 +1,348 @@ +# Module-based Model Runtime Interface for AOT + +- Feature Name: modu

[GitHub] [tvm] Mousius commented on a change in pull request #10095: [mircoTVM] Include standalone_crt dependencies in MLF

Mousius commented on a change in pull request #10095: URL: https://github.com/apache/tvm/pull/10095#discussion_r794436250 ## File path: apps/microtvm/ethosu/Makefile ## @@ -19,13 +19,13 @@ # Setup build environment BUILD_DIR := build -STANDALONE_CRT_PATH := $(shell python3

[GitHub] [tvm] grant-arm opened a new pull request #10095: [mircoTVM] Include standalone_crt dependencies in MLF

grant-arm opened a new pull request #10095: URL: https://github.com/apache/tvm/pull/10095 This PR includes the standalone_crt files in the MLF tarball if the runtime is `crt` . It also updates the `metadata.json` file as described in the ["Capturing dependent libraries..." discussion](

[tvm] branch ci-docker-staging updated (96ef1d0 -> a45c68d)

This is an automated email from the ASF dual-hosted git repository. mousius pushed a change to branch ci-docker-staging in repository https://gitbox.apache.org/repos/asf/tvm.git. discard 96ef1d0 Test new Docker images omit e6af874 [PyTorch] Fix rsub type (#10090) omit fa317ed [CI][Fi

[GitHub] [tvm] vvchernov commented on pull request #9980: [VirtualMachine] fix raw pointer using by VirtualMachine

vvchernov commented on pull request #9980: URL: https://github.com/apache/tvm/pull/9980#issuecomment-1024104646 Thanks @driazati ! It works -- 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 sp

[GitHub] [tvm] argrento commented on pull request #9811: PackedFunction to return params from the .so module, show warning when no params are set

argrento commented on pull request #9811: URL: https://github.com/apache/tvm/pull/9811#issuecomment-1024100624 Thank you for the comments! Currently I have 2 solutions: 1. Restore the original logic of Run function 2. Since current time complexity of Run is O(N^2), I can rewrite it and

[GitHub] [tvm] lhutton1 commented on a change in pull request #10084: Upgrade tensorflow to version to 2.6.x

lhutton1 commented on a change in pull request #10084: URL: https://github.com/apache/tvm/pull/10084#discussion_r794374573 ## File path: docker/install/ubuntu_install_tensorflow.sh ## @@ -21,6 +21,5 @@ set -u set -o pipefail pip3 install \ -"h5py<3.0" \ Review comment:

[GitHub] [tvm] masahi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

masahi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794377291 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1769,10 +1780,27 @@ def get_v(v, default_v): return v return default_

[GitHub] [tvm] masahi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

masahi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794377291 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1769,10 +1780,27 @@ def get_v(v, default_v): return v return default_

[GitHub] [tvm] Mousius opened a new pull request #10094: [microTVM] Update Zephyr to 2.7

Mousius opened a new pull request #10094: URL: https://github.com/apache/tvm/pull/10094 This supports the reference system added in #9853 -- 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 spec

[GitHub] [tvm] Mousius commented on pull request #10093: [CI] Move docs precheck to docs: GPU

Mousius commented on pull request #10093: URL: https://github.com/apache/tvm/pull/10093#issuecomment-1024057823 CC @areusch @leandron -- 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

[tvm] branch ci-docker-staging updated (134edbd -> 96ef1d0)

This is an automated email from the ASF dual-hosted git repository. mousius pushed a change to branch ci-docker-staging in repository https://gitbox.apache.org/repos/asf/tvm.git. discard 134edbd Test new docker images add d97274c [MetaSchedule] Post Processor: Rewrite Reduction Block (#

[tvm] 01/01: Test new Docker images

// NOTE: these lines are scanned by docker/dev_common.sh. Please update the regex as needed. --> -ci_lint = "tlcpack/ci-lint:v0.67" -ci_gpu = "tlcpack/ci-gpu:v0.80" -ci_cpu = "tlcpack/ci-cpu:v0.80" +ci_lint = "tlcpackstaging/ci_lint:20220128-070420-fa317

[GitHub] [tvm] Mousius opened a new pull request #10093: [CI] Move docs precheck to docs: GPU

Mousius opened a new pull request #10093: URL: https://github.com/apache/tvm/pull/10093 This didn't trigger in #10077 as it lives in unittest that is skipped on docs only builds. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to

[GitHub] [tvm] t-vi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

t-vi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794349364 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1769,10 +1780,27 @@ def get_v(v, default_v): return v return default_v

[GitHub] [tvm] t-vi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

t-vi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794349364 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1769,10 +1780,27 @@ def get_v(v, default_v): return v return default_v

[GitHub] [tvm] masahi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

masahi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794341021 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1769,10 +1780,27 @@ def get_v(v, default_v): return v return default_

[GitHub] [tvm] masahi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

masahi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794340375 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1755,7 +1765,8 @@ def pad(inputs, input_types): return pad -def clamp(self, inputs

[GitHub] [tvm] MasterJH5574 commented on a change in pull request #10092: [MetaSchedule][M4a] Mutator: Mutate-Tile-Size

MasterJH5574 commented on a change in pull request #10092: URL: https://github.com/apache/tvm/pull/10092#discussion_r794320827 ## File path: tests/python/unittest/test_meta_schedule_mutator_mutate_tile_size.py ## @@ -0,0 +1,92 @@ +# Licensed to the Apache Software Foundation (A

[GitHub] [tvm] Hzfengsy commented on a change in pull request #10092: [MetaSchedule][M4a] Mutator: Mutate-Tile-Size

Hzfengsy commented on a change in pull request #10092: URL: https://github.com/apache/tvm/pull/10092#discussion_r794319461 ## File path: tests/python/unittest/test_meta_schedule_mutator_mutate_tile_size.py ## @@ -0,0 +1,92 @@ +# Licensed to the Apache Software Foundation (ASF)

[GitHub] [tvm] MasterJH5574 opened a new pull request #10092: [MetaSchedule][M4a] Mutator: Mutate-Tile-Size

MasterJH5574 opened a new pull request #10092: URL: https://github.com/apache/tvm/pull/10092 This PR is part of the Meta-Schedule upstreaming effort (#8473), which adds the tile size mutator. cc @junrushao1994 @spectrometerHBH @Hzfengsy @comaniac --- Co-authored-by: Ju

[GitHub] [tvm] KennyTang1988 commented on pull request #9655: [Caffe Frontend] Add support for Power layer

KennyTang1988 commented on pull request #9655: URL: https://github.com/apache/tvm/pull/9655#issuecomment-1023995316 > @fernchen @FrozenGene Can you take a look at this PR, please? Thanks! I hava update commit, please help check it again. -- This is an automated message from the Apa

[GitHub] [tvm] t-vi commented on a change in pull request #10091: [Torch] Experimental support for FX-quantized models

t-vi commented on a change in pull request #10091: URL: https://github.com/apache/tvm/pull/10091#discussion_r794290423 ## File path: python/tvm/relay/frontend/pytorch.py ## @@ -1755,7 +1765,8 @@ def pad(inputs, input_types): return pad -def clamp(self, inputs,

[tvm] branch main updated (fa317ed -> e6af874)

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 fa317ed [CI][Fix] Remove additional qnn.op.transpose_conv2d from docs (#10083) add e6af874 [PyTorch] Fix rsub type

[GitHub] [tvm] masahi merged pull request #10090: [PyTorch] Fix rsub type

masahi merged pull request #10090: URL: https://github.com/apache/tvm/pull/10090 -- 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.