Meteorix commented on a change in pull request #7147:
URL: https://github.com/apache/tvm/pull/7147#discussion_r547098688
##
File path: python/tvm/topi/nn/batch_matmul.py
##
@@ -59,3 +60,25 @@ def batch_matmul(x, y, oshape=None):
lambda b, i, j: te.sum(x[b if XB != 1 el
kongroo commented on a change in pull request #7147:
URL: https://github.com/apache/tvm/pull/7147#discussion_r547096507
##
File path: python/tvm/topi/nn/batch_matmul.py
##
@@ -59,3 +60,25 @@ def batch_matmul(x, y, oshape=None):
lambda b, i, j: te.sum(x[b if XB != 1 els
Meteorix commented on a change in pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#discussion_r547093078
##
File path: python/tvm/topi/cuda/batch_matmul_tensorcore.py
##
@@ -0,0 +1,274 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more c
Meteorix commented on a change in pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#discussion_r547091793
##
File path: python/tvm/relay/op/strategy/cuda.py
##
@@ -657,6 +657,20 @@ def batch_matmul_strategy_cuda(attrs, inputs, out_type,
target):
nam
Meteorix commented on a change in pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#discussion_r547091707
##
File path: python/tvm/topi/cuda/batch_matmul_tensorcore.py
##
@@ -0,0 +1,274 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more c
ANSHUMAN87 commented on a change in pull request #7107:
URL: https://github.com/apache/tvm/pull/7107#discussion_r547091469
##
File path: tests/scripts/task_ci_python_setup.sh
##
@@ -31,3 +31,4 @@ set -o pipefail
echo "Addtiional setup in" ${CI_IMAGE_NAME}
python3 -m pip ins
jcf94 commented on pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#issuecomment-749364296
@Meteorix, great thanks for your PR! The code looks good to me.
This is an automated message from the Apache Git Servi
ANSHUMAN87 opened a new pull request #7148:
URL: https://github.com/apache/tvm/pull/7148
This is a follow up PR.
1. It has resolved the issue in CSR scheduling for both Cuda & X86.
2. Also the test cases in Tensorflow frontends are enabled for the same.
cc @tkonolige , @FrozenG
jcf94 commented on a change in pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#discussion_r547085365
##
File path: python/tvm/relay/op/strategy/cuda.py
##
@@ -657,6 +657,20 @@ def batch_matmul_strategy_cuda(attrs, inputs, out_type,
target):
name="
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r547087287
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,52 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_in
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r547086641
##
File path: src/relay/op/tensor/transform.cc
##
@@ -1553,6 +1553,52 @@ RELAY_REGISTER_OP("meshgrid")
.set_attr("FTVMCompute", MeshgridCompute)
Meteorix commented on a change in pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#discussion_r547086764
##
File path: python/tvm/topi/cuda/batch_matmul_tensorcore.py
##
@@ -0,0 +1,275 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more c
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r547086590
##
File path: python/tvm/relay/op/_transform.py
##
@@ -63,6 +63,8 @@
_reg.register_injective_schedule("sparse_to_dense")
_reg.register_injective_schedu
Meteorix opened a new pull request #7147:
URL: https://github.com/apache/tvm/pull/7147
Add legalize pass: padding dense/conv2d/batch_matmul ops to legal shapes for
using tensorcore on cuda target. To limit the overhead introduced by padding,
we count the `extra_flops` and set the threshold
jcf94 commented on a change in pull request #7146:
URL: https://github.com/apache/tvm/pull/7146#discussion_r547081523
##
File path: python/tvm/topi/cuda/batch_matmul_tensorcore.py
##
@@ -0,0 +1,275 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more cont
Meteorix opened a new pull request #7146:
URL: https://github.com/apache/tvm/pull/7146
Add batch_matmul tensorcore schedule for bert inference. It shows better
performance than cublas batch_matmul kernel.
@jcf94 @merrymercy could you help review this pr?
---
FrozenGene commented on pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#issuecomment-749329511
Thanks everyone @comaniac @jcf94 @giuseros
This is an automated message from the Apache Git Service.
To respond
This is an automated email from the ASF dual-hosted git repository.
zhaowu pushed a change to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git.
from 24cd869 [COMMUNITY] @jcf94 -> Committer (#7141)
add 98ca771 [Auto Scheduler] Mali Support (#7132)
No new revisions
FrozenGene merged pull request #7132:
URL: https://github.com/apache/tvm/pull/7132
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 t
comaniac commented on a change in pull request #7145:
URL: https://github.com/apache/tvm/pull/7145#discussion_r547055396
##
File path: python/tvm/auto_scheduler/search_task.py
##
@@ -221,10 +221,6 @@ def __init__(
target_host = Target(target_host)
self.d
comaniac commented on a change in pull request #7145:
URL: https://github.com/apache/tvm/pull/7145#discussion_r547055396
##
File path: python/tvm/auto_scheduler/search_task.py
##
@@ -221,10 +221,6 @@ def __init__(
target_host = Target(target_host)
self.d
jcf94 commented on a change in pull request #7145:
URL: https://github.com/apache/tvm/pull/7145#discussion_r547038245
##
File path: python/tvm/auto_scheduler/search_task.py
##
@@ -221,10 +221,6 @@ def __init__(
target_host = Target(target_host)
self.dag
comaniac opened a new pull request #7145:
URL: https://github.com/apache/tvm/pull/7145
The previous implementation of task serialization mechanism has a potential
problem, which is a bit tricky so I didn't notice that before.
Previously, I defined a set of attributes in the Python ob
comaniac opened a new pull request #7144:
URL: https://github.com/apache/tvm/pull/7144
Current record processing APIs in auto_scheduler are based on files. This PR
adds two APIs to manipulate tuning records in memory.
cc @merrymercy @jcf94
-
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 9713d67 Created CSourceMetaData module for model metadata (#7002)
add 24cd869 [COMMUNITY] @jcf94 -> Committer (#
junrushao1994 merged pull request #7141:
URL: https://github.com/apache/tvm/pull/7141
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 t
codeislife99 commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546961924
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_in
comaniac opened a new pull request #7143:
URL: https://github.com/apache/tvm/pull/7143
The current auto_scheduler callbacks can only be implemented in C++. This PR
exposes the interface of measure callbacks to Python so that people can plug in
their own callback functions easily.
No
zhiics commented on pull request #7002:
URL: https://github.com/apache/tvm/pull/7002#issuecomment-749223654
Thanks @manupa-arm @areusch
This is an automated message from the Apache Git Service.
To respond to the message, ple
zhiics merged pull request #7002:
URL: https://github.com/apache/tvm/pull/7002
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 s
This is an automated email from the ASF dual-hosted git repository.
zhic 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 9713d67 Created CSourceMetaData module for model metadat
Wheest opened a new pull request #7142:
URL: https://github.com/apache/tvm/pull/7142
The goal of this pull request is to make asymmetric padding a first-class
citizen in 2D convolution in TOPI.
The current workload description has `"hpad"` and `"wpad"`, however this is
not represent
kevinthesun commented on a change in pull request #7137:
URL: https://github.com/apache/tvm/pull/7137#discussion_r546923370
##
File path: python/tvm/relay/frontend/pytorch.py
##
@@ -1857,16 +1857,18 @@ def nms(self, inputs, input_types):
scores = inputs[1]
iou
kevinthesun commented on a change in pull request #7137:
URL: https://github.com/apache/tvm/pull/7137#discussion_r546923370
##
File path: python/tvm/relay/frontend/pytorch.py
##
@@ -1857,16 +1857,18 @@ def nms(self, inputs, input_types):
scores = inputs[1]
iou
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 82942fb [TOPI] Simplify GPU NMS IR and optimize a bit (#7136)
add 53c0641 Fix a few OpNode argument field descr
mbrookhart commented on pull request #7140:
URL: https://github.com/apache/tvm/pull/7140#issuecomment-749189788
Thanks @domin1985 @junrushao1994
This is an automated message from the Apache Git Service.
To respond to the mes
mbrookhart merged pull request #7140:
URL: https://github.com/apache/tvm/pull/7140
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 t
mbrookhart commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546916764
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_inde
codeislife99 commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546895484
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_in
codeislife99 commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546893836
##
File path: python/tvm/relay/op/_transform.py
##
@@ -63,6 +63,7 @@
_reg.register_injective_schedule("sparse_to_dense")
_reg.register_injective_schedu
tkonolige commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546891906
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_index
Wheest edited a comment on pull request #6137:
URL: https://github.com/apache/tvm/pull/6137#issuecomment-749136716
Hello there, updating this pull request to be up-to-date with the latest
`main` branch.
In terms of things remaining to do:
- [x] [Consider using compute_at and v
codeislife99 commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546888428
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_in
codeislife99 commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546888139
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_in
Wheest edited a comment on pull request #6137:
URL: https://github.com/apache/tvm/pull/6137#issuecomment-749136716
Hello there, updating this pull request to be up-to-date with the latest
`main` branch.
In terms of things remaining to do:
- [x] [Consider using compute_at and v
Wheest edited a comment on pull request #6137:
URL: https://github.com/apache/tvm/pull/6137#issuecomment-749136716
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub an
Wheest edited a comment on pull request #6137:
URL: https://github.com/apache/tvm/pull/6137#issuecomment-749136716
Hello there, updating this pull request to be up-to-date with the latest
`main` branch.
In terms of things remaining to do:
- [x] [Consider using compute_at and v
Wheest commented on pull request #6137:
URL: https://github.com/apache/tvm/pull/6137#issuecomment-749136716
Hello there, updating this pull request to be up-to-date with the latest
`main` branch.
In terms of things remaining to do:
- [x] [Consider using compute_at and vectoriz
TylerADavis commented on a change in pull request #7134:
URL: https://github.com/apache/tvm/pull/7134#discussion_r546871956
##
File path: tests/python/frontend/pytorch/test_forward.py
##
@@ -1889,21 +1889,73 @@ def _get_default_vm_targets():
return [tgt for (tgt, _) in tvm
TylerADavis commented on a change in pull request #7134:
URL: https://github.com/apache/tvm/pull/7134#discussion_r546871956
##
File path: tests/python/frontend/pytorch/test_forward.py
##
@@ -1889,21 +1889,73 @@ def _get_default_vm_targets():
return [tgt for (tgt, _) in tvm
tqchen opened a new pull request #7141:
URL: https://github.com/apache/tvm/pull/7141
Please join us to welcome @jcf94 as a new committer. He is one of the major
contributors to the latest AutoScheduler(AutoTVMv2). He also actively supports
the community on
answering questions related t
tkonolige commented on a change in pull request #7107:
URL: https://github.com/apache/tvm/pull/7107#discussion_r546840492
##
File path: tests/scripts/task_ci_python_setup.sh
##
@@ -31,3 +31,4 @@ set -o pipefail
echo "Addtiional setup in" ${CI_IMAGE_NAME}
python3 -m pip inst
tkonolige commented on a change in pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#discussion_r546836442
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,84 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_index
tkonolige commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r546831211
##
File path: python/tvm/relay/op/_transform.py
##
@@ -63,6 +63,8 @@
_reg.register_injective_schedule("sparse_to_dense")
_reg.register_injective_schedule(
Laurawly commented on pull request #7136:
URL: https://github.com/apache/tvm/pull/7136#issuecomment-749093671
Thanks @masahi @mbrookhart
This is an automated message from the Apache Git Service.
To respond to the message, pl
Laurawly merged pull request #7136:
URL: https://github.com/apache/tvm/pull/7136
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
This is an automated email from the ASF dual-hosted git repository.
laurawly 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 82942fb [TOPI] Simplify GPU NMS IR and optimize a bi
manupa-arm commented on pull request #7002:
URL: https://github.com/apache/tvm/pull/7002#issuecomment-749003721
Ping @areusch @zhiics.
Let me know if you have any more concerns, I ll be disappearing for holidays
from tomorrow.
Thus, would like to address if there is anything today. T
tqchen commented on pull request #7127:
URL: https://github.com/apache/tvm/pull/7127#issuecomment-748981949
Thanks @hogepodge !
This is an automated message from the Apache Git Service.
To respond to the message, please log o
tqchen commented on pull request #7093:
URL: https://github.com/apache/tvm/pull/7093#issuecomment-748982915
Thanks @euntaik @giuseros @FrozenGene !
This is an automated message from the Apache Git Service.
To respond to the m
tqchen merged pull request #7093:
URL: https://github.com/apache/tvm/pull/7093
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 s
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/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new 9914685 [TFLite] add support for float16 (#7093)
99146
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/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new 38273ee [CUDA] Parallel Cuda Mergesort (#7099)
38273ee
tqchen merged pull request #7099:
URL: https://github.com/apache/tvm/pull/7099
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 s
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/tvm.git
The following commit(s) were added to refs/heads/main by this push:
new bf5e248 Added additional information to the from_onnx
tqchen merged pull request #7127:
URL: https://github.com/apache/tvm/pull/7127
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 s
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r546677613
##
File path: include/tvm/topi/transform.h
##
@@ -1386,6 +1386,85 @@ inline Array meshgrid(const Array&
inputs, const std::string& in
return result;
FrozenGene commented on pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#issuecomment-748942493
@comaniac @jcf94 @giuseros have addressed the comments. Please have another
one round of look.
This is an automa
FrozenGene commented on a change in pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#discussion_r546664255
##
File path: python/tvm/relay/op/strategy/mali.py
##
@@ -69,6 +71,36 @@ def conv2d_strategy_mali(attrs, inputs, out_type, target):
raise
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r546652775
##
File path: python/tvm/relay/op/transform.py
##
@@ -1320,3 +1320,47 @@ def adv_index(inputs):
Output tensor.
"""
return _make.adv_in
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r546652731
##
File path: python/tvm/topi/transform.py
##
@@ -931,3 +931,47 @@ def adv_index(data, indices):
Output tensor
"""
return cpp.adv_inde
codeislife99 commented on a change in pull request #7125:
URL: https://github.com/apache/tvm/pull/7125#discussion_r546652595
##
File path: include/tvm/topi/transform.h
##
@@ -1386,6 +1386,85 @@ inline Array meshgrid(const Array&
inputs, const std::string& in
return result;
codeislife99 edited a comment on pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#issuecomment-748921727
cc : @trevor-m @comaniac @zhiics PTAL
This is an automated message from the Apache Git Service.
To res
codeislife99 commented on pull request #7126:
URL: https://github.com/apache/tvm/pull/7126#issuecomment-748921727
cc : @trevor-m @comaniac @zhiics
This is an automated message from the Apache Git Service.
To respond to the
FrozenGene commented on a change in pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#discussion_r546625878
##
File path: python/tvm/topi/nn/conv2d.py
##
@@ -382,7 +382,15 @@ def conv2d_nhwc(
if auto_scheduler_rewritten_layout:
# Infer shape for the
FrozenGene commented on a change in pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#discussion_r546625280
##
File path: src/auto_scheduler/search_task.cc
##
@@ -90,6 +90,22 @@ HardwareParams
HardwareParamsNode::GetDefaultHardwareParams(const Target& target
FrozenGene commented on a change in pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#discussion_r546625280
##
File path: src/auto_scheduler/search_task.cc
##
@@ -90,6 +90,22 @@ HardwareParams
HardwareParamsNode::GetDefaultHardwareParams(const Target& target
FrozenGene commented on a change in pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#discussion_r546624511
##
File path: python/tvm/topi/mali/conv2d.py
##
@@ -545,6 +577,19 @@ def _alter_conv2d_layout(attrs, inputs, tinfos, out_type):
return None
+@
domin1985 opened a new pull request #7140:
URL: https://github.com/apache/tvm/pull/7140
Fix a few missed OpNode argument field descriptions when registered.
@mbrookhart @junrushao1994
This is an automated message
giuseros commented on a change in pull request #7132:
URL: https://github.com/apache/tvm/pull/7132#discussion_r546591992
##
File path: python/tvm/topi/nn/conv2d.py
##
@@ -382,7 +382,15 @@ def conv2d_nhwc(
if auto_scheduler_rewritten_layout:
# Infer shape for the r
masahi opened a new issue #7139:
URL: https://github.com/apache/tvm/issues/7139
https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/PR-7136/8/pipeline
This is an automated message from the Apache Git Service.
To resp
81 matches
Mail list logo