lixiaoquan opened a new pull request #6704:
URL: https://github.com/apache/incubator-tvm/pull/6704
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL
areusch commented on pull request #6703:
URL: https://github.com/apache/incubator-tvm/pull/6703#issuecomment-710744702
also, vagrant boxes currently hosted at:
https://app.vagrantup.com/areusch/boxes/microtvm-staging
This
areusch opened a new pull request #6703:
URL: https://github.com/apache/incubator-tvm/pull/6703
This PR adds two Vagrantfiles:
- a µTVM base box in `tools/microtvm/base-box` intended to support general
µTVM development. it includes all the dependencies necessary to build the
Zephyr
tqchen opened a new pull request #6702:
URL: https://github.com/apache/incubator-tvm/pull/6702
cc @yzhliu @ZihengJiang
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
sxjscience commented on a change in pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#discussion_r506775666
##
File path: python/tvm/relay/frontend/mxnet.py
##
@@ -627,6 +632,21 @@ def _mx_expand_dims(inputs, attrs):
return
sxjscience commented on a change in pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#discussion_r506769946
##
File path: python/tvm/topi/x86/batch_matmul.py
##
@@ -157,6 +163,10 @@ def batch_matmul_cblas(cfg, x, y):
YB, N, YK =
comaniac commented on a change in pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#discussion_r506769623
##
File path: python/tvm/topi/x86/batch_matmul.py
##
@@ -157,6 +163,10 @@ def batch_matmul_cblas(cfg, x, y):
YB, N, YK =
tkonolige commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506766614
##
File path: cmake/modules/RustExt.cmake
##
@@ -0,0 +1,26 @@
+if(USE_RUST_EXT AND NOT USE_RUST_EXT EQUAL OFF)
Review comment:
I think
tkonolige commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506766711
##
File path: cmake/modules/RustExt.cmake
##
@@ -0,0 +1,26 @@
+if(USE_RUST_EXT AND NOT USE_RUST_EXT EQUAL OFF)
Review comment:
tkonolige commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506766711
##
File path: cmake/modules/RustExt.cmake
##
@@ -0,0 +1,26 @@
+if(USE_RUST_EXT AND NOT USE_RUST_EXT EQUAL OFF)
Review comment:
_if()
jroesch commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506766408
##
File path: cmake/modules/RustExt.cmake
##
@@ -0,0 +1,26 @@
+if(USE_RUST_EXT AND NOT USE_RUST_EXT EQUAL OFF)
Review comment:
I am not a
sxjscience commented on a change in pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#discussion_r506766247
##
File path: python/tvm/topi/x86/batch_matmul.py
##
@@ -157,6 +163,10 @@ def batch_matmul_cblas(cfg, x, y):
YB, N, YK =
rkimball commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506764888
##
File path: cmake/modules/RustExt.cmake
##
@@ -0,0 +1,26 @@
+if(USE_RUST_EXT AND NOT USE_RUST_EXT EQUAL OFF)
Review comment:
We really
jroesch commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506762857
##
File path: rust/compiler-ext/src/lib.rs
##
@@ -0,0 +1,43 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more
jroesch commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506762875
##
File path: rust/compiler-ext/src/lib.rs
##
@@ -0,0 +1,43 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more
tkonolige commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506762570
##
File path: rust/tvm/src/ir/diagnostics/codespan.rs
##
@@ -0,0 +1,183 @@
+use std::collections::HashMap;
Review comment:
This file
tkonolige commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506761995
##
File path: rust/compiler-ext/src/lib.rs
##
@@ -0,0 +1,43 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more
jroesch commented on a change in pull request #6656:
URL: https://github.com/apache/incubator-tvm/pull/6656#discussion_r506761896
##
File path: rust/compiler-ext/Cargo.toml
##
@@ -0,0 +1,16 @@
+[package]
+name = "compiler-ext"
+version = "0.1.0"
+authors = ["Jared Roesch "]
sxjscience commented on pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#issuecomment-710693179
The integration tests take a very long time because there are two many
combinations. For example:
tkonolige opened a new pull request #6701:
URL: https://github.com/apache/incubator-tvm/pull/6701
This dependency is needed for #6679. I also had to add a cmake build from
source to the i386 image so that xgboost builds (it builds from source because
there is no wheel for i386).
jwfromm opened a new pull request #6700:
URL: https://github.com/apache/incubator-tvm/pull/6700
This PR adds a converter for Loop operators in Onnx. Loops in onnx are
represented as entire onnx graphs embedded within the op. To support this
structure, there are few changes to the
jwfromm commented on pull request #6700:
URL: https://github.com/apache/incubator-tvm/pull/6700#issuecomment-710681034
@masahi @mbrookhart @csullivan @soiferj Can you guys help review this PR?
This is an automated message
jroesch commented on pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#issuecomment-710677184
As we add more tests can we measure what kind of time increase this will
induce in CI? integration tests are becoming increasingly slow and expensive to
run. cc @areusch
tqchen commented on a change in pull request #6692:
URL: https://github.com/apache/incubator-tvm/pull/6692#discussion_r506732411
##
File path: include/tvm/ir/diagnostic_context.h
##
@@ -18,71 +18,25 @@
*/
/*!
- * \file diagnostic.h
+ * \file diagnostic_context.h
Review
jroesch merged pull request #6698:
URL: https://github.com/apache/incubator-tvm/pull/6698
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
This is an automated email from the ASF dual-hosted git repository.
jroesch pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from e997185 [Relay] Change some passes to mix mode (#6695)
add 461e75b [LLVM][WINDOWS] Recover windows
comaniac commented on a change in pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#discussion_r506718116
##
File path: python/tvm/relay/frontend/mxnet.py
##
@@ -58,6 +58,11 @@
_activation_map = {"sigmoid": _op.sigmoid, "tanh": _op.tanh, "relu":
rkimball commented on pull request #6698:
URL: https://github.com/apache/incubator-tvm/pull/6698#issuecomment-710586184
I tested with llvm 11.0 release and this PR and my failing unit test now
passes. Thank you @tqchen for this.
sxjscience commented on pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699#issuecomment-710579653
@yzhliu @comaniac @icemelon9
This is an automated message from the Apache Git Service.
To respond to
sxjscience opened a new pull request #6699:
URL: https://github.com/apache/incubator-tvm/pull/6699
Fix the MXNet 2.0 integration in relay. Tested the BERT and ALBERT model in
the new [GluonNLP 1.0](https://github.com/dmlc/gluon-nlp/tree/master) and has
passed the test. I will later add
sxjscience closed pull request #6696:
URL: https://github.com/apache/incubator-tvm/pull/6696
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
tkonolige commented on a change in pull request #6685:
URL: https://github.com/apache/incubator-tvm/pull/6685#discussion_r506601559
##
File path: tests/python/frontend/tensorflow/test_forward.py
##
@@ -1750,6 +1750,64 @@ def test_forward_batch_matmul():
tqchen commented on pull request #6698:
URL: https://github.com/apache/incubator-tvm/pull/6698#issuecomment-710202863
cc @tmoreau89 @rkimball
This is an automated message from the Apache Git Service.
To respond to the
tqchen opened a new pull request #6698:
URL: https://github.com/apache/incubator-tvm/pull/6698
Windows COFF requires comdat information to support weak-like linkage(via
any).
This patch fixes the windows LLVM support after LLVM-8.
tqchen edited a comment on pull request #6692:
URL: https://github.com/apache/incubator-tvm/pull/6692#issuecomment-710126019
Thanks @rkimball . I feel the PR can still be simplified further to avoid
potential confusion of two diagnostic header files(one minimum and another
built for IR).
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/main by this push:
new e997185 [Relay] Change some passes to mix
tqchen merged pull request #6695:
URL: https://github.com/apache/incubator-tvm/pull/6695
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
tqchen commented on pull request #6692:
URL: https://github.com/apache/incubator-tvm/pull/6692#issuecomment-710126019
As a minimum set of changes to the PR. i would recommend keep the files as
they are, and just move the ICHECK defs into `support/logging.h`
the additional
tqchen commented on a change in pull request #6692:
URL: https://github.com/apache/incubator-tvm/pull/6692#discussion_r506560441
##
File path: include/tvm/support/diagnostic_context.h
##
@@ -27,62 +27,20 @@
* replace the existing errors.h.
*/
-#ifndef
tqchen commented on a change in pull request #6692:
URL: https://github.com/apache/incubator-tvm/pull/6692#discussion_r506559623
##
File path: include/tvm/support/diagnostic.h
##
@@ -0,0 +1,74 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more
mbrookhart commented on pull request #6695:
URL: https://github.com/apache/incubator-tvm/pull/6695#issuecomment-710104990
@jroesch or @tqchen, care to take a look and/or merge?
This is an automated message from the Apache
mbrookhart commented on a change in pull request #6695:
URL: https://github.com/apache/incubator-tvm/pull/6695#discussion_r506527342
##
File path: src/relay/transforms/fold_constant.cc
##
@@ -118,7 +120,7 @@ class ConstantFolder : public ExprMutator {
}
}
- Expr
mbaret commented on pull request #6697:
URL: https://github.com/apache/incubator-tvm/pull/6697#issuecomment-710020641
cc @manupa-arm @zhiics @comaniac
This is an automated message from the Apache Git Service.
To respond to
mbaret opened a new pull request #6697:
URL: https://github.com/apache/incubator-tvm/pull/6697
Currently, all codegens using BYOC must make use of the default
ConstantUpdater pass. However, certain codegens, like Ethos-N, don't want to
store any constants in metadata module. This provides
hzfan commented on issue #6691:
URL: https://github.com/apache/incubator-tvm/issues/6691#issuecomment-709956681
Sure. I will look into that. Thanks @trevor-m
This is an automated message from the Apache Git Service.
To
masahi merged pull request #6602:
URL: https://github.com/apache/incubator-tvm/pull/6602
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
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/incubator-tvm.git.
from 4c4d3dc Fix tutorial broken by Docker build (#6694)
add cf49e8b [Torch, Quantization] Necessary
masahi commented on pull request #6602:
URL: https://github.com/apache/incubator-tvm/pull/6602#issuecomment-709954612
Thanks @siju-samuel
This is an automated message from the Apache Git Service.
To respond to the message,
sxjscience opened a new pull request #6696:
URL: https://github.com/apache/incubator-tvm/pull/6696
Fix the MXNet 2.0 integration in relay. Tested the BERT and ALBERT model in
the new GluonNLP v1 and has passed the test. Will later on add unittests in
GluonNLP to ensure that most backbones
49 matches
Mail list logo
