tqchen commented on pull request #5415:
URL: https://github.com/apache/incubator-tvm/pull/5415#issuecomment-618767898
Good catch, dump_ir is temporary disabled because we can move to the new
PassContext trace API. We should remove that line, @wpan11nv can you send a
patch?
tqchen commented on pull request #5426:
URL: https://github.com/apache/incubator-tvm/pull/5426#issuecomment-618771249
cc @zhiics @icemelon9 @jroesch
This is a PR that helps to prepare the std::string->String migration.
tqchen opened a new pull request #5426:
URL: https://github.com/apache/incubator-tvm/pull/5426
To make runtime.String to work as naturally as possible in the python side,
we make it sub-class the python's str object. Note that however, we cannot
sub-class Object at the same time due
This is an automated email from the ASF dual-hosted git repository.
haichen pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from 6c77195 [FRONTEND][MXNET] support elemwise logic ops (#5361)
add e68450d [PY][FFI] Introduce
wpan11nv commented on pull request #5415:
URL: https://github.com/apache/incubator-tvm/pull/5415#issuecomment-618741145
The following line is broken after this change:
https://github.com/apache/incubator-tvm/blob/master/python/tvm/driver/build_module.py#L161
tqchen commented on issue #5047:
URL: https://github.com/apache/incubator-tvm/issues/5047#issuecomment-618772966
https://github.com/apache/incubator-tvm/pull/5426
This is an automated message from the Apache Git Service.
To
tqchen edited a comment on pull request #5426:
URL: https://github.com/apache/incubator-tvm/pull/5426#issuecomment-618771249
cc @zhiics @icemelon9 @jroesch @wweic
This is a PR that helps to prepare the std::string->String migration.
yongfeng-nv commented on a change in pull request #5367:
URL: https://github.com/apache/incubator-tvm/pull/5367#discussion_r414259311
##
File path: include/tvm/arith/int_set.h
##
@@ -152,6 +152,22 @@ class IntSet : public ObjectRef {
yongfeng-nv commented on a change in pull request #5367:
URL: https://github.com/apache/incubator-tvm/pull/5367#discussion_r414263234
##
File path: src/arith/int_set.cc
##
@@ -311,6 +311,21 @@ inline IntervalSet Combine(Analyzer*
analyzer,
LOG(FATAL) << "Modular by
michalpiszczek commented on pull request #5425:
URL: https://github.com/apache/incubator-tvm/pull/5425#issuecomment-618720491
PTAL
This is an automated message from the Apache Git Service.
To respond to the message,
tmoreau89 commented on pull request #5425:
URL: https://github.com/apache/incubator-tvm/pull/5425#issuecomment-618721380
Thanks @michalpiszczek @tqchen, the PR has been merged
This is an automated message from the Apache Git
This is an automated email from the ASF dual-hosted git repository.
moreau pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from 6faacc6 [MXNET]DepthToSpace & SpaceToDepth Operator (#5408)
add cf5c63b Add option to specify
tqchen commented on pull request #5361:
URL: https://github.com/apache/incubator-tvm/pull/5361#issuecomment-618773375
Thanks @kazum @maheshambule
This is an automated message from the Apache Git Service.
To respond to the
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/master by this push:
new 6c77195 [FRONTEND][MXNET] support
tqchen commented on pull request #5345:
URL: https://github.com/apache/incubator-tvm/pull/5345#issuecomment-618773499
please suggest a resolution(move to a new folder) and act on it :)
This is an automated message from the
yongfeng-nv commented on a change in pull request #5367:
URL: https://github.com/apache/incubator-tvm/pull/5367#discussion_r414259311
##
File path: include/tvm/arith/int_set.h
##
@@ -152,6 +152,22 @@ class IntSet : public ObjectRef {
icemelon9 commented on pull request #5426:
URL: https://github.com/apache/incubator-tvm/pull/5426#issuecomment-618803899
Thanks @tqchen @zhiics
This is an automated message from the Apache Git Service.
To respond to the
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618732386
With https://github.com/dmlc/xgboost/pull/5590, I can now run
`tests/python/unittest/test_autotvm_xgboost_model.py::test_fit` without
crashing.
woniuasd commented on issue #5133:
URL: https://github.com/apache/incubator-tvm/issues/5133#issuecomment-618812554
my pytorch code is "canvas[:, indices] = voxels"
The following operators are not implemented: ['aten::index_put_']
tqchen commented on pull request #4805:
URL: https://github.com/apache/incubator-tvm/pull/4805#issuecomment-618773667
ping
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618348796
@hcho3 I found the cause:
The `dmlc::Error` is from `libtvm.so` instead of `libxgboost.so`, which can
not be caught by XGBoost. I believe there are some mixed
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618368767
Somehow the `dmlc::LogMessageFatal` constructor is a plt function, but the
one in tvm is not plt, and calling it in XGBoost got resolved into the one in
tvm.
```
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618368767
Somehow the `dmlc::LogMessageFatal` constructor is a plt function in
XGBoost, but the one in tvm is not plt, and calling it in XGBoost got resolved
into the one in
JonathanMace opened a new issue #5420:
URL: https://github.com/apache/incubator-tvm/issues/5420
I'm training a resnet18_v2 model for CUDA using the RPC runner. I run an
RPC server on a remote machine that has a GPU, and I run the training program
locally on a machine that has no GPUs.
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618348796
@hcho3 I found the cause:
The `dmlc::Error` is from `libtvm.so` instead of `libxgboost.so`, which can
not be caught by XGBoost. I believe there's some mixed
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618348796
@hcho3 I found the cause:
The dmlc error is thrown from `libtvm.so` instead of `libxgboost.so`, which
can not be caught by XGBoost. I believe there's some mixed up
leandron commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618373787
> Somehow the `dmlc::LogMessageFatal` constructor is a plt function in
XGBoost, but the one in tvm is not plt, and calling it in XGBoost got resolved
into the one in tvm.
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618350152
But I don't understand how it happens.
This is an automated message from the Apache Git Service.
To respond
kazum commented on a change in pull request #5331:
URL: https://github.com/apache/incubator-tvm/pull/5331#discussion_r413546386
##
File path: python/tvm/relay/op/nn/_nn.py
##
@@ -458,6 +458,15 @@ def compute_cross_entropy(attrs, inputs, out_dtype):
srkreddy1238 commented on issue #5404:
URL: https://github.com/apache/incubator-tvm/issues/5404#issuecomment-618227043
cc @jroesch any clues here ?
This is an automated message from the Apache Git Service.
To respond to the
kazum opened a new pull request #5419:
URL: https://github.com/apache/incubator-tvm/pull/5419
Fixes sphinx3 warnings, #5396.
This also merge redundant SDAccel and AOCL docs into one as `hls.rst`.
@tqchen @tmoreau89 Please help to review.
This is an automated email from the ASF dual-hosted git repository.
zhaowu pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/master by this push:
new 1acad98 [RUNTIME][CONTRIB] CoreML
FrozenGene commented on issue #5283:
URL: https://github.com/apache/incubator-tvm/pull/5283#issuecomment-618234736
Thanks @kazum Wish we could have one end2end tutorial next, especially how
to leverage iPhone NPU Engine. I think this will attract many people,
especially for Mobile
mbaret commented on issue #5409:
URL: https://github.com/apache/incubator-tvm/pull/5409#issuecomment-618290965
The motivation for this is that when you have a supported operator, such as
qnn.concatenate, which takes some constant inputs, those constant nodes are
annotated as 'default'. So
siju-samuel commented on issue #5394:
URL: https://github.com/apache/incubator-tvm/pull/5394#issuecomment-618468746
@inadob Thanks. I will give a try.
This is an automated message from the Apache Git Service.
To respond to
u99127 commented on issue #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#issuecomment-618493747
>
>
Thanks for the quick review.
> It would be great if we can avoid the hack into the `with_same_user`. One
alternative would be still pass in the
notoraptor commented on a change in pull request #5331:
URL: https://github.com/apache/incubator-tvm/pull/5331#discussion_r413952517
##
File path: python/tvm/relay/op/nn/_nn.py
##
@@ -458,6 +458,15 @@ def compute_cross_entropy(attrs, inputs, out_dtype):
lhutton1 commented on issue #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#issuecomment-618515093
cc @anijain2305
This is an automated message from the Apache Git Service.
To respond to the message, please
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618406401
@hcho3
Could you please create another build with position independent code of
dmlc-core disabled?
```
set_target_properties(dmlc PROPERTIES
inadob commented on issue #5394:
URL: https://github.com/apache/incubator-tvm/pull/5394#issuecomment-618459856
@siju-samuel
- Have you tried to recreate the TFL operations using
tf.quantization.quantize()
u99127 commented on issue #5394:
URL: https://github.com/apache/incubator-tvm/pull/5394#issuecomment-618499720
>
>
> @inadob Thanks. I will give a try.
I have a more fundamental question - I don't expect Quantize and Dequantize
to show up in Tflite models for inference as
tom-gall commented on issue #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#issuecomment-618499668
> It would be great if we can avoid the hack into the `with_same_user`. One
alternative would be still pass in the `PYTEST_ADDOPTS` env variable from the
docker env(for
anijain2305 commented on issue #4803:
URL: https://github.com/apache/incubator-tvm/pull/4803#issuecomment-618524196
Thanks @FrozenGene @tqchen @inadob
This is merged!
This is an automated message from the Apache Git
This is an automated email from the ASF dual-hosted git repository.
anijain2305 pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from dbd0114 fix [RUNTIME][VULKAN] vkBuffer released before memory copy
command send to GPU (#5388) (#5418)
u99127 commented on a change in pull request #5330:
URL: https://github.com/apache/incubator-tvm/pull/5330#discussion_r413943986
##
File path: python/tvm/relay/frontend/tflite.py
##
@@ -1067,6 +1069,27 @@ def convert_zeros_like(self, op):
return out
+def
tqchen commented on issue #5367:
URL: https://github.com/apache/incubator-tvm/pull/5367#issuecomment-618503288
@hzfan please take another look and
https://tvm.apache.org/docs/contribute/code_review.html#approve-and-request-changes-explicitly
u99127 commented on issue #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#issuecomment-618504136
>
>
> > It would be great if we can avoid the hack into the `with_same_user`.
One alternative would be still pass in the `PYTEST_ADDOPTS` env variable from
the docker
lhutton1 opened a new pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422
* This patch means that you can specify an additional layout, rather than
using the layout chosen by default during conversion.
* This is specifically useful for external codegen when a 3rd
tqchen commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618472379
I wonder if that was due to inconsistency between dmlc-core of tvm and xgb.
https://github.com/apache/incubator-tvm/pull/5401 updated the logging to
latest, please check again.
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r413969744
##
File path: include/tvm/relay/op_attr_types.h
##
@@ -158,13 +158,15 @@ using FTVMAlterOpLayout = runtime::TypedPackedFunc<
* \param tinfos
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618385906
@leandron I believe it's the root cause of this issue. I just don't know
how to fix it. Maybe it's a bug in system linker or compiler used the build to
pip
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618406401
@hcho3
Could you please create another build with position independent code of
dmlc-core disabled?
```
set_target_properties(dmlc PROPERTIES
u99127 commented on issue #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#issuecomment-618416038
I don't like my current hack of overloading "with_same_user" for sourcing
this global environment but it seemed like the simplest hack and worked in my
environment. Obviously I
u99127 edited a comment on issue #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#issuecomment-618416038
I don't like my current hack of overloading "with_same_user" for sourcing
this global environment but it seemed like the simplest hack and worked in my
environment.
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/master by this push:
new 708fd9a [DOCS] Migrate HLS documents
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618385906
@leandron I believe it's the root cause of this issue. I just don't know
how to fix it. Maybe it's a bug in system linker on 18.04 or compiler used to
build the
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618385906
@leandron I believe it's the root cause of this issue. I just don't know
how to fix it. Maybe it's a bug in system linker 18.04 or compiler used to
build the pip
tqchen commented on issue #5420:
URL: https://github.com/apache/incubator-tvm/issues/5420#issuecomment-618456921
This is something that we could improve as part of the target system
https://discuss.tvm.ai/t/target-and-attributes/6013 they should be part of the
target attributes instead of
tqchen commented on issue #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#issuecomment-618455854
It would be great if we can avoid the hack into the `with_same_user`. One
alternative would be still pass in the `PYTEST_ADDOPTS` env variable from the
docker env(for development
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618406401
@hcho3
Could you please create another build with position independent code of
dmlc-core disabled?
```
set_target_properties(dmlc PROPERTIES
yongfeng-nv commented on a change in pull request #5367:
URL: https://github.com/apache/incubator-tvm/pull/5367#discussion_r413843721
##
File path: src/arith/int_set.cc
##
@@ -372,7 +387,27 @@ class IntervalSetEvaluator :
}
IntervalSet Eval(const PrimExpr& val) {
-
yongfeng-nv commented on a change in pull request #5367:
URL: https://github.com/apache/incubator-tvm/pull/5367#discussion_r413843881
##
File path: src/arith/int_set.cc
##
@@ -518,7 +533,14 @@ class IntervalSetEvaluator :
// whether set is exactly single point that equals
tqchen commented on issue #5419:
URL: https://github.com/apache/incubator-tvm/pull/5419#issuecomment-618449524
Thanks @kazum !
This is an automated message from the Apache Git Service.
To respond to the message, please log
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from 708fd9a [DOCS] Migrate HLS documents from md to rst (#5419)
add dbd0114 fix [RUNTIME][VULKAN] vkBuffer
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618385906
@leandron I believe it's the root cause of this issue. I just don't know
how to fix it. Maybe it's a bug in system linker or compiler used to build to
pip package? Or a
u99127 opened a new pull request #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421
In many places having a global pytest flag is useful . For me with the
build and test of tvm , I would like to be able to globally pass in
pytest options as part of development flow or CI
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618385906
@leandron I believe it's the root cause of this issue. I just don't know
how to fix it. Maybe it's a bug in system linker or compiler used to build the
pip
comaniac edited a comment on issue #5409:
URL: https://github.com/apache/incubator-tvm/pull/5409#issuecomment-618537434
Ah I see your point. I missed considering that `ConstantNode` is also a
"node" instead of a "var". I agree with you that constant nodes should follow
the target of the
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r414011046
##
File path: include/tvm/relay/op_attr_types.h
##
@@ -158,13 +158,15 @@ using FTVMAlterOpLayout = runtime::TypedPackedFunc<
* \param tinfos
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618560244
@trivialfis Here it is:
https://drive.google.com/file/d/13WZRRaUPKil4rwH2avgUix_xO5LpIYs5/view?usp=sharing
This
u99127 commented on pull request #5392:
URL: https://github.com/apache/incubator-tvm/pull/5392#issuecomment-618582877
Gentle ping .
This is an automated message from the Apache Git Service.
To respond to the message, please
u99127 commented on a change in pull request #5421:
URL: https://github.com/apache/incubator-tvm/pull/5421#discussion_r414036151
##
File path: docs/contribute/pull_request.rst
##
@@ -118,3 +118,6 @@ If you want to run a single test:
rm -rf python/tvm/*.pyc
wpan11nv opened a new pull request #5424:
URL: https://github.com/apache/incubator-tvm/pull/5424
- remove unnecessary white spaces from storage kind
- do not start a new scope for vectorization as temporary
variables are alll uniquely generated.
The above two changes make
wpan11nv edited a comment on pull request #5424:
URL: https://github.com/apache/incubator-tvm/pull/5424#issuecomment-618591146
Notably extra white spaces or scopes are removed. When the loop nesting or
expression is deep, the emitted becomes unreadable.
Before:
`extern "C"
wpan11nv edited a comment on pull request #5424:
URL: https://github.com/apache/incubator-tvm/pull/5424#issuecomment-618591146
Notably extra white spaces or scopes are removed. When the loop nesting or
expression is deep, the emitted becomes unreadable.
Before:
`extern "C"
wpan11nv edited a comment on pull request #5424:
URL: https://github.com/apache/incubator-tvm/pull/5424#issuecomment-618591146
Notably extra white spaces or scopes are removed. When the loop nesting or
expression is deep, the emitted becomes unreadable.
Before:
`extern "C"
This is an automated email from the ASF dual-hosted git repository.
haichen pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from a3b1397 [Frontend] Asymmetric padding of convolution support (#4803)
add 9c12ec8 [cuDNN] Add cuDNN
michalpiszczek commented on pull request #5425:
URL: https://github.com/apache/incubator-tvm/pull/5425#issuecomment-618603153
@tmoreau89
This is an automated message from the Apache Git Service.
To respond to the message,
michalpiszczek opened a new pull request #5425:
URL: https://github.com/apache/incubator-tvm/pull/5425
Adds an option to `config.cmake` for specifying the location of FlatBuffers
when building with TFLite ON.
This option _must_ be set when building full TVM with TFLite ON, but is
tqchen commented on pull request #5423:
URL: https://github.com/apache/incubator-tvm/pull/5423#issuecomment-618627418
The inhertitance will still work for 3rdparty, because overflow is enabled
in all these cases, the support path will be like the same before this PR. we
could discuss
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618642163
@hcho3 I tried master branch and the commit before:
> I wonder if that was due to inconsistency between dmlc-core of tvm and
xgb. #5401 updated the logging to latest,
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618645218
@hcho3 Yes. Currently detached at the commit before above linked PR.
```
fis@fis-Standard-PC-Q35-ICH9-2009:~/Workspace/XGBoost/incubator-tvm$ git
status
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618645218
@hcho3 Yes.
```
fis@fis-Standard-PC-Q35-ICH9-2009:~/Workspace/XGBoost/incubator-tvm$ git
status
HEAD detached at 56941fb9d
Untracked files:
(use "git
trivialfis edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618640842
@hcho3 One last request, otherwise I'm running out of ideas.
Patch both xgboost and rabit's C API macro:
For xgboost:
``` patch
diff --git
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618649425
@areusch Can you try out the latest TVM master on your end? I'm still having
trouble reproducing.
This is an
hcho3 edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618649425
@areusch Can you try out the latest TVM master on your end? I'm still having
trouble reproducing the original issue.
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618656005
Finally, I reproduced it. Yes!
This is an automated message from the Apache Git Service.
To respond to the
u99127 edited a comment on pull request #5394:
URL: https://github.com/apache/incubator-tvm/pull/5394#issuecomment-618499720
>
>
> @inadob Thanks. I will give a try.
I have a more fundamental question - I don't expect Quantize and Dequantize
to show up in Tflite models for
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/master by this push:
new 1f6c498 [DOCS] Migrate some markdowns
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from d81a4fa [CI] Migrate Tensorflow and Tensorflow lite in CI to 2.1.0
(#5392)
add 1f6c498 [DOCS]
tqchen edited a comment on pull request #5423:
URL: https://github.com/apache/incubator-tvm/pull/5423#issuecomment-618627418
The inhertitance will still work for 3rdparty, because overflow is enabled
in all of these cases, the support path will be like the same before this PR.
we could
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618640842
@hcho3 One last request, otherwise I'm running out of ideas.
Patch both xgboot and rabit's C API macro:
For xgboost:
``` patch
diff --git
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618641439
@trivialfis Did you update TVM to latest?
This is an automated message from the Apache Git Service.
To respond
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618644587
@trivialfis And you ran `git submodule update --init --recursive`?
This is an automated message from the Apache
jroesch commented on pull request #5144:
URL: https://github.com/apache/incubator-tvm/pull/5144#issuecomment-618651343
@icemelon9 yes my intention is we can in theory dynamically compute
everything like malloc or pool allocators.
hcho3 edited a comment on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618656005
Finally, I reproduced it. Yes! Note: I used latest TVM as of today. The
crash still occured.
This is an
trivialfis commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618661450
@hcho3 Could you try applying the patches I posted above and use the
corresponding cmake flags?
This is an
tqchen commented on issue #5423:
URL: https://github.com/apache/incubator-tvm/pull/5423#issuecomment-618535184
cc @jroesch @zhiics @wweic @junrushao1994 @yzhliu
This is an automated message from the Apache Git Service.
To
tqchen opened a new pull request #5423:
URL: https://github.com/apache/incubator-tvm/pull/5423
The _type_child_slots can be used to enable quick type checking optimization
by checking the whether the type index is within the bound.
This PR enables these static slots:
-
tqchen commented on issue #5423:
URL: https://github.com/apache/incubator-tvm/pull/5423#issuecomment-618535594
current summary of the type table(non leaf nodes)
```
$python -c "import tvm; tvm.runtime._ffi_api.DumpTypeTable(1)"
[9] BaseExpr parent=runtime.Object
1 - 100 of 150 matches
Mail list logo
