anijain2305 commented on issue #4307: [QNN][Legalize] Specialize for Platforms
w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#issuecomment-553274508
Addressed the comments :)
This is an automated
optima2005 commented on a change in pull request #4300:
[Relay][Frontend][Tensorflow]Add conv2d_transpose
URL: https://github.com/apache/incubator-tvm/pull/4300#discussion_r345595368
##
File path: tests/python/frontend/tensorflow/test_forward.py
##
@@ -349,6 +361,14 @@ def
junrushao1994 commented on issue #4297: [TOPI][OP] Support Faster-RCNN Proposal
OP on CPU
URL: https://github.com/apache/incubator-tvm/pull/4297#issuecomment-553265084
@were Could you also look into the hybrid script issue?
T
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345585786
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345586389
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345584658
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345583599
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345588091
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345584117
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
jackwish commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345587849
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124 @@
FrozenGene commented on a change in pull request #4307: [QNN][Legalize]
Specialize for Platforms w/o fast Int8 support
URL: https://github.com/apache/incubator-tvm/pull/4307#discussion_r345584407
##
File path: python/tvm/relay/qnn/op/legalizations.py
##
@@ -137,4 +165,124
This is an automated email from the ASF dual-hosted git repository.
wuwei pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from e541c75 Fix the TF tutorial to run against TF2.0 and TF1.x (#4104)
add 8cd5cce [TOPI][OP] Support Faster
vinx13 commented on issue #4297: [TOPI][OP] Support Faster-RCNN Proposal OP on
CPU
URL: https://github.com/apache/incubator-tvm/pull/4297#issuecomment-553255882
Thanks @FrozenGene @kevinthesun this is merged
This is an automa
vinx13 merged pull request #4297: [TOPI][OP] Support Faster-RCNN Proposal OP on
CPU
URL: https://github.com/apache/incubator-tvm/pull/4297
This is an automated message from the Apache Git Service.
To respond to the message,
vinx13 commented on a change in pull request #4297: [TOPI][OP] Support
Faster-RCNN Proposal OP on CPU
URL: https://github.com/apache/incubator-tvm/pull/4297#discussion_r345583973
##
File path: topi/python/topi/vision/rcnn/proposal.py
##
@@ -60,6 +61,261 @@ def reg_iou(x1,
FrozenGene commented on a change in pull request #4297: [TOPI][OP] Support
Faster-RCNN Proposal OP on CPU
URL: https://github.com/apache/incubator-tvm/pull/4297#discussion_r345583061
##
File path: topi/python/topi/vision/rcnn/proposal.py
##
@@ -60,6 +61,261 @@ def reg_iou(
FrozenGene opened a new pull request #4326: [TFLite] Fix Prelu unified shape
error
URL: https://github.com/apache/incubator-tvm/pull/4326
Our slope shape needs to be 1D.
cc: @apivovarov Could you try this fix?
Thi
FrozenGene removed a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553205441
@apivovarov I think change the axis from 3 to -1 should resolve your
problem, because we want to act on Channel. if we set -1, it could know the
wweic opened a new pull request #4325: [Prelude][Relay] Add get tensor data
utilities
URL: https://github.com/apache/incubator-tvm/pull/4325
This implements issue https://github.com/apache/incubator-tvm/issues/4291.
I also applied the optimization in tensorflow converter that it keep
vinx13 commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose
output_padding parameter
URL: https://github.com/apache/incubator-tvm/pull/4318#issuecomment-553221068
We need to update logs in
https://github.com/uwsampl/tvm-distro/tree/master/tophub
-
vinx13 edited a comment on issue #4318: [Relay][TOPI]Fix meaning of
conv2d_transpose output_padding parameter
URL: https://github.com/apache/incubator-tvm/pull/4318#issuecomment-553220724
there are some issue with ci, please retrigger the ci
vinx13 commented on issue #4318: [Relay][TOPI]Fix meaning of conv2d_transpose
output_padding parameter
URL: https://github.com/apache/incubator-tvm/pull/4318#issuecomment-553220724
there are some issue with ci, please retriever the ci
---
FrozenGene commented on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553205441
@apivovarov I think change the axis from 3 to -1 should resolve your
problem, because we want to act on Channel. if we set -1, it could know the
channel
FrozenGene commented on issue #4320: Fix TFLite RESHAPE assert
URL: https://github.com/apache/incubator-tvm/pull/4320#issuecomment-553201791
could we add one unittest?
This is an automated message from the Apache Git Service.
FrozenGene commented on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553201683
Ok. You could refer my unittest and see how the prelu is constructed,
because tf doesn’t have prelu, tflite just recognize op pattern to produce
prelu.
alexgl-github closed pull request #4322: Fix conv2d_transpose compilation bug
for CUDA.
URL: https://github.com/apache/incubator-tvm/pull/4322
This is an automated message from the Apache Git Service.
To respond to the messa
alexgl-github commented on issue #4322: Fix conv2d_transpose compilation bug
for CUDA.
URL: https://github.com/apache/incubator-tvm/pull/4322#issuecomment-553200969
Closing for now because of different direction to solve this
apivovarov commented on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553199610
I can not share the file unfortunately. I'll try to create similar tflite
file to reproduce the issue.
-
FrozenGene commented on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553197887
could you share your custom tflite model of prelu? Seems that we should
reshape alpha tensor to 4D make share meet TVM requirement.
TaoLv commented on a change in pull request #4323: [Contrib] Add MKL DNN option
URL: https://github.com/apache/incubator-tvm/pull/4323#discussion_r345533163
##
File path: cmake/modules/contrib/BLAS.cmake
##
@@ -55,3 +55,11 @@ elseif(USE_BLAS STREQUAL "none")
else()
mess
icemelon9 opened a new pull request #4324: [Relay] Use opaque for where op
URL: https://github.com/apache/incubator-tvm/pull/4324
Fuse where op will cause significant performance degradation in BERT model
This is an automated
icemelon9 opened a new pull request #4323: [Contrib] Add MKL DNN option
URL: https://github.com/apache/incubator-tvm/pull/4323
Thanks for contributing to TVM! Please refer to guideline
https://docs.tvm.ai/contribute/ for useful information and tips. After the pull
request is submitted, p
alexgl-github opened a new pull request #4322: Fix conv2d_transpose compilation
bug for CUDA.
URL: https://github.com/apache/incubator-tvm/pull/4322
Disable fallback schedule for 1xN (N>1) kernels.
Thanks for contributing to TVM! Please refer to guideline
https://docs.tvm.ai/contr
yzhliu commented on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate
URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-553183479
@kevinthesun Thanks. I'll go through the merging history make sure
everything since the last monthly report gets into the release note.
-
kevinthesun commented on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate
URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-553178514
Symbolic Shape Enhancement:
Add more shape functions commonly used in cv models:
https://github.com/apache/incubator-tvm/pull/4179
yzhliu merged pull request #4104: Fix the TF tutorial to run against TF2.0 and
TF1.x
URL: https://github.com/apache/incubator-tvm/pull/4104
This is an automated message from the Apache Git Service.
To respond to the message,
yzhliu commented on issue #4104: Fix the TF tutorial to run against TF2.0 and
TF1.x
URL: https://github.com/apache/incubator-tvm/pull/4104#issuecomment-553171416
Thanks @ic @yongwww @tqchen
This is an automated message from
This is an automated email from the ASF dual-hosted git repository.
liuyizhi pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from 03a29da [Relay][Op][TF] Complete tensor array unstack with all ranks
support (#4309)
add e541c75 Fix
petrex opened a new pull request #4321: Enable hipModuleGetGlobal() in ROCm
module
URL: https://github.com/apache/incubator-tvm/pull/4321
Enable hipModuleGetGlobal() to returns device memory pointer and size of the
kernel.
--
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553159880
PRELU docs says that H and W can be "shared". That is what we see in my
custom tflite model.
PRELU Alpha tensor with index 4 has shape [1,1,16
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553159880
PRELU docs says that H and W can be "shared". That is what we see in my
custom tflite model.
PRELU Alpha tensor with index 4 has shape [1,1,16
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553159880
PRELU docs says that H and W can be shared. That is what we see in my custom
tflite model.
PRELU Alpha tensor with index 4 has shape [1,1,16].
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553159880
PRELU docs says that H and W can be "shared". That is what we see in my
custom tflite model.
PRELU Alpha tensor with index 4 has shape [1,1,16
apivovarov commented on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553159880
PRELU docs says that H and W can be shared. That is what we see in my custom
tflite model.
PRELU Alpha tensor with index 4 has shape [1,1,16]. It mea
zhiics edited a comment on issue #4280: [TVM][RUNTIME] A minimum example to
generate external library wrappers for DSOModule
URL: https://github.com/apache/incubator-tvm/pull/4280#issuecomment-553143166
yes, we don't need to call this as a runtime now. It actually just uses
DSOModule and C
yongwww commented on a change in pull request #4300:
[Relay][Frontend][Tensorflow]Add conv2d_transpose
URL: https://github.com/apache/incubator-tvm/pull/4300#discussion_r345487638
##
File path: tests/python/frontend/tensorflow/test_forward.py
##
@@ -349,6 +361,14 @@ def te
weberlo commented on a change in pull request #4274: [µTVM] Enable AutoTVM for
ARM STM32F746XX Boards
URL: https://github.com/apache/incubator-tvm/pull/4274#discussion_r345485880
##
File path: src/runtime/micro/micro_session.cc
##
@@ -76,79 +214,122 @@ MicroSession::~Micro
apivovarov commented on issue #4320: Fix TFLite RESHAPE assert
URL: https://github.com/apache/incubator-tvm/pull/4320#issuecomment-553149893
@FrozenGene Can you have a look?
This is an automated message from the Apache Git Se
apivovarov opened a new pull request #4320: Fix TFLite RESHAPE assert
URL: https://github.com/apache/incubator-tvm/pull/4320
Recently I found that Reshape op input might have one or two input tensors.
in TVM code we assert that length is two but we only use `input_tensor[0]`.
Looks lik
soiferj commented on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553147128
Thanks a lot for the investigation - that's a useful finding. There should
be no harm in adding this include :)
-
jmorrill commented on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553146605
Ok thats it. `string` pulls in `cctype` in VS2019 and does not in 2017
Following builds in 2019 and does not in 201
jmorrill edited a comment on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553142604
I compiled file `\tvm\nnvm\src\top\tensor\transform.cc` with `/showIncludes`
because it uses `std::isdigit` and fou
yzhliu edited a comment on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate
URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-553143216
Kindly remind that we would like to cut the release branch as well as the RC
tag (RC0) this week. If you want to get any feature in
yzhliu edited a comment on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate
URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-553143216
Kindly remind that we would like to cut the release branch as well as the RC
tag (RC0) this week. If you want to get any feature in
yzhliu commented on issue #4259: [DEV][DRAFT] TVM v0.6 Release candidate
URL: https://github.com/apache/incubator-tvm/issues/4259#issuecomment-553143216
Kindly remind that we would like to cut the release branch as well as the RC
tag (RC0) this week. If you want to get any feature into the
zhiics commented on issue #4280: [TVM][RUNTIME] A minimum example to generate
external library wrappers for DSOModule
URL: https://github.com/apache/incubator-tvm/pull/4280#issuecomment-553143166
yes, we don't need to call this as a runtime now. It actually just use
DSOModule and CSourceMo
jmorrill edited a comment on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553142604
I compiled file `\tvm\nnvm\src\top\tensor\transform.cc` with `/showIncludes`
because it uses `std::isdigit` and fou
jmorrill commented on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553142604
I compiled file `\tvm\nnvm\src\top\tensor\transform.cc` with `/showIncludes`
because it uses `std::isdigit `and found the
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553132406
@FrozenGene
I tried to compile my model and got the following `unable to unify` errors.
( I also added model visualization screenshots at the
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553132406
@FrozenGene
I tried to compile my model and got the following `unable to unify` errors.
( I also added model visualization screenshots at the
comaniac commented on issue #4258: [WIP][TVM] Bring Your Own Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#issuecomment-553136026
@u99127 thanks for th comments.
As we are making another PR #4280 for the runtime and will refine this PR
accordingly after #4280
jmorrill edited a comment on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553135764
> I'm using VS 2017 (cl version 19.16.27032.1). Where does your compiler get
`std::isdigit` from?
I believe
jmorrill commented on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553135764
> I'm using VS 2017 (cl version 19.16.27032.1). Where does your compiler get
`std::isdigit` from?
I believe mine is
yongwww commented on a change in pull request #4308: [Frontend]Add TensorFlow
FloorMod
URL: https://github.com/apache/incubator-tvm/pull/4308#discussion_r345466063
##
File path: python/tvm/relay/op/tensor.py
##
@@ -397,6 +397,24 @@ def divide(lhs, rhs):
return _make.d
comaniac commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345466726
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache So
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553132406
@FrozenGene
I tried to compile my model and got the following `unable to unify` errors.
( I also added model visualization screenshots at the
comaniac commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345465735
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache So
apivovarov edited a comment on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553132406
@FrozenGene
I tried to compile my model and got the following `unable to unify` errors
```
%0 = nn.pad(%input_1, pad_width=[[0, 0], [0,
tqchen commented on issue #4280: [TVM][RUNTIME] A minimum example to generate
external library wrappers for DSOModule
URL: https://github.com/apache/incubator-tvm/pull/4280#issuecomment-553132659
Thanks @comaniac i feel that given it is already unified with the DSOModule,
we don't have to
apivovarov commented on issue #4298: [TFLite] Support PRelu
URL: https://github.com/apache/incubator-tvm/pull/4298#issuecomment-553132406
@FrozenGene
I tried to compile my model and got the following errors
```
%0 = nn.pad(%input_1, pad_width=[[0, 0], [0, 1], [0, 1], [0, 0]]);
soiferj commented on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553131671
I'm using VS 2017 (cl version 19.16.27032.1). Where does your compiler get
`std::isdigit` from?
I think we should st
jmorrill commented on issue #4319: [Build][Windows] Fix Windows build by
including cctype
URL: https://github.com/apache/incubator-tvm/pull/4319#issuecomment-553129461
With the latest source from master, I do not get any build errors on the
CMake generated tvm.sln.
I am using Visu
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345448945
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345447964
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345448945
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345447589
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345448429
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345444040
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345445780
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
u99127 commented on a change in pull request #4258: [WIP][TVM] Bring Your Own
Codegen to TVM
URL: https://github.com/apache/incubator-tvm/pull/4258#discussion_r345447170
##
File path: tutorials/dev/custom_relay_backend.py
##
@@ -0,0 +1,291 @@
+# Licensed to the Apache Soft
kevinthesun commented on a change in pull request #4312: [TOPI][Relay][OP]
Dynamic NMS and strided_slice
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r345404746
##
File path: include/tvm/relay/attrs/vision.h
##
@@ -96,6 +96,7 @@ struct GetValidCountsAt
petrex commented on issue #4305: Proper Device Attribute Query for AMD GPU
URL: https://github.com/apache/incubator-tvm/pull/4305#issuecomment-553117279
The error doesn't seem relevant to me, let me kick off the CI again.
```
___ test_forward_placeholder
_
kevinthesun commented on a change in pull request #4312: [TOPI][Relay][OP]
Dynamic NMS and strided_slice
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r345446273
##
File path: src/relay/op/vision/nms.cc
##
@@ -136,8 +143,9 @@
TVM_REGISTER_API("relay.op
kevinthesun commented on a change in pull request #4312: [TOPI][Relay][OP]
Dynamic NMS and strided_slice
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r345445628
##
File path: src/relay/op/vision/nms.cc
##
@@ -107,6 +112,7 @@ bool NMSRel(const Array& ty
kevinthesun commented on a change in pull request #4312: [TOPI][Relay][OP]
Dynamic NMS and strided_slice
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r345444532
##
File path: src/relay/op/tensor/transform.cc
##
@@ -1947,57 +1948,81 @@ Array > StridedSl
soiferj opened a new pull request #4319: [Build][Windows] Fix Windows build
URL: https://github.com/apache/incubator-tvm/pull/4319
The Windows build is currently broken with the error: `isdigit is not a
member of standard`. It looks like `std::isdigit` is in header ``.
Including it fixes t
kevinthesun commented on a change in pull request #4312: [TOPI][Relay][OP]
Dynamic NMS and strided_slice
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r345442538
##
File path: python/tvm/relay/op/vision/nms.py
##
@@ -55,6 +55,7 @@ def get_valid_counts(d
kevinthesun commented on a change in pull request #4312: [TOPI][Relay][OP]
Dynamic NMS and strided_slice
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r345442353
##
File path: python/tvm/relay/op/vision/_vision.py
##
@@ -99,23 +100,34 @@ def schedule_nm
abergeron opened a new pull request #4318: [Relay][TOPI]Fix meaning of
conv2d_transpose output_padding parameter
URL: https://github.com/apache/incubator-tvm/pull/4318
I've fixed the meaning of output_padding to be more in line with what other
machine learning frameworks and libraries inte
kevinthesun commented on issue #4308: [Frontend]Add TensorFlow FloorMod
URL: https://github.com/apache/incubator-tvm/pull/4308#issuecomment-553110700
@tqchen @soiferj Updated.
This is an automated message from the Apache Git S
This is an automated email from the ASF dual-hosted git repository.
zhic pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from e680611 Add test for the qnn_add operator (#4282)
add 03a29da [Relay][Op][TF] Complete tensor array unsta
zhiics commented on issue #4309: [Relay][Op][TF] Complete tensor array unstack
with all ranks support
URL: https://github.com/apache/incubator-tvm/pull/4309#issuecomment-553102978
Thanks everyone. This is now merged.
This is
zhiics merged pull request #4309: [Relay][Op][TF] Complete tensor array unstack
with all ranks support
URL: https://github.com/apache/incubator-tvm/pull/4309
This is an automated message from the Apache Git Service.
To respo
kevinthesun edited a comment on issue #4308: [Frontend]Add TensorFlow FloorMod
URL: https://github.com/apache/incubator-tvm/pull/4308#issuecomment-553099510
OK. I'll add topi floordiv/mod, which calls lower floor div/mod if dtype is
int/uint, and combinations of other topi op if float. Do y
This is an automated email from the ASF dual-hosted git repository.
zhic pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from dddb0ed add (#4311)
add e680611 Add test for the qnn_add operator (#4282)
No new revisions were added by
u99127 commented on issue #4292: Retain qnn input kernel scales
URL: https://github.com/apache/incubator-tvm/pull/4292#issuecomment-55309
All updates now done ,please review and merge as appropriate .
This is an automated
kevinthesun commented on issue #4308: [Frontend]Add TensorFlow FloorMod
URL: https://github.com/apache/incubator-tvm/pull/4308#issuecomment-553099510
OK. I'll add topi floordiv/mod, which calls lower floor div/mod if dtype is
int/uint, and combinations of other topi op if float.
--
soiferj commented on a change in pull request #4309: [Relay][Op][TF] Complete
tensor array unstack with all ranks support
URL: https://github.com/apache/incubator-tvm/pull/4309#discussion_r345426740
##
File path: python/tvm/relay/frontend/tensorflow.py
##
@@ -1070,9 +1071,
wweic opened a new pull request #4317: [Test][Relay][Pass] Add test case for
lambda lift
URL: https://github.com/apache/incubator-tvm/pull/4317
cc @jroesch @icemelon9 @zhiics @MarisaKirisame
This is an automated message from
tqchen commented on issue #4308: [Frontend]Add TensorFlow FloorMod
URL: https://github.com/apache/incubator-tvm/pull/4308#issuecomment-553090157
yes, for technical simplicity low level floordiv/mod only support integer,
but you can always float into floor(div()) in topi
--
1 - 100 of 135 matches
Mail list logo
