junrushao1994 commented on pull request #5910:
URL: https://github.com/apache/incubator-tvm/pull/5910#issuecomment-648592923
Hmmm could you briefly describe what it is? Thanks!
This is an automated message from the Apache Git
This is an automated email from the ASF dual-hosted git repository.
sijusamuel pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from e8ccfd0 Allow implicit conversion in TVM FFI to tvm::Bool (#5907)
add 4c78c03 PyTorch frontend: fix
siju-samuel merged pull request #5897:
URL: https://github.com/apache/incubator-tvm/pull/5897
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
jinboci opened a new pull request #5910:
URL: https://github.com/apache/incubator-tvm/pull/5910
Thanks for contributing to TVM! Please refer to guideline
https://tvm.apache.org/docs/contribute/ for useful information and tips. After
the pull request is submitted, please request code revi
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 a1fb841 Add Binary Intrinsic ops to TIR Ops in C++ (#5900)
add e8ccfd0 Allow implicit conversion in TVM
tqchen merged pull request #5907:
URL: https://github.com/apache/incubator-tvm/pull/5907
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 g
jcf94 commented on pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#issuecomment-648550481
> Please split the rewrite simplify into its own PR and add unittestcases
for the newly added rules. It would be great if we can write down a few lines
of proof sketches, to
yzhliu merged pull request #5906:
URL: https://github.com/apache/incubator-tvm/pull/5906
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 g
This is an automated email from the ASF dual-hosted git repository.
liuyizhi pushed a commit to branch v0.6
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/v0.6 by this push:
new c03d97c [Fix] Fix conv2d alter op for arm
yzhliu merged pull request #5904:
URL: https://github.com/apache/incubator-tvm/pull/5904
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 g
This is an automated email from the ASF dual-hosted git repository.
liuyizhi pushed a change to branch v0.6
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from 804b7fa Fix dense (#4728) (#5905)
add 3cd42c9 [Relay][Fix] Fix alter op layout when calling a global va
yzhliu merged pull request #5905:
URL: https://github.com/apache/incubator-tvm/pull/5905
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 g
This is an automated email from the ASF dual-hosted git repository.
liuyizhi pushed a commit to branch v0.6
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/v0.6 by this push:
new 804b7fa Fix dense (#4728) (#5905)
804b7fa
icemelon9 merged pull request #5903:
URL: https://github.com/apache/incubator-tvm/pull/5903
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 t
This is an automated email from the ASF dual-hosted git repository.
haichen pushed a commit to branch v0.6
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/v0.6 by this push:
new 85d99e7 [Relay][Pass] Fix lambda lift pass
comaniac opened a new pull request #5909:
URL: https://github.com/apache/incubator-tvm/pull/5909
Per discussion in
https://discuss.tvm.ai/t/pattern-matching-for-tuplegetitem/7069, this PR
supports arbitrary index matching for tuple get item node.
Document is also updated.
cc
tqchen closed issue #5908:
URL: https://github.com/apache/incubator-tvm/issues/5908
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
tqchen commented on issue #5908:
URL: https://github.com/apache/incubator-tvm/issues/5908#issuecomment-648508016
Thanks @BelfortXP Please open a new trouble shooting thread on
https://discuss.tvm.ai/
This is an automated m
BelfortXP commented on issue #5908:
URL: https://github.com/apache/incubator-tvm/issues/5908#issuecomment-648505381
These are the information in spike, you can see this if the image cannot be
loaded.
spike --rbb-port=9824 -m0x1000:0x2 rot13-64
Listening for remote bitbang
BelfortXP removed a comment on issue #5908:
URL: https://github.com/apache/incubator-tvm/issues/5908#issuecomment-648502621
![image](https://user-images.githubusercontent.com/39782968/85481184-ed3b7a80-b5f3-11ea-9489-c08ec2a2279e.png)
--
BelfortXP commented on issue #5908:
URL: https://github.com/apache/incubator-tvm/issues/5908#issuecomment-648502621
![image](https://user-images.githubusercontent.com/39782968/85481184-ed3b7a80-b5f3-11ea-9489-c08ec2a2279e.png)
--
BelfortXP opened a new issue #5908:
URL: https://github.com/apache/incubator-tvm/issues/5908
Anybody knows how to use of the riscv-spike? The strange thing is that the
spike shows "get unsupported command" when execute the test_alloc() in
tvm\tests\python\unittest\test_runtime_micro.py.
tqchen commented on pull request #5900:
URL: https://github.com/apache/incubator-tvm/pull/5900#issuecomment-648476992
Thanks @mbrookhart @junrushao1994 !
This is an automated message from the Apache Git Service.
To respond to
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 81e8cfc add a few gradients (#5899)
add a1fb841 Add Binary Intrinsic ops to TIR Ops in C++ (#5900)
No
tqchen merged pull request #5900:
URL: https://github.com/apache/incubator-tvm/pull/5900
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 g
junrushao1994 opened a new pull request #5907:
URL: https://github.com/apache/incubator-tvm/pull/5907
This allows conversion to tvm::Bool from both int and tvm objects in TVM FFI.
CC: @mbrookhart @antinucleon @jwfromm
trevor-m commented on pull request #5857:
URL: https://github.com/apache/incubator-tvm/pull/5857#issuecomment-648450759
@ wpan11nv Any more comments?
This is an automated message from the Apache Git Service.
To respond to the
trevor-m edited a comment on pull request #5857:
URL: https://github.com/apache/incubator-tvm/pull/5857#issuecomment-648450759
@wpan11nv Any more comments?
This is an automated message from the Apache Git Service.
To respond
tmoreau89 commented on pull request #9:
URL: https://github.com/apache/incubator-tvm-vta/pull/9#issuecomment-648436019
That is correct, we'll need to bump the versioning of the bitstream
This is an automated message from the
icemelon9 opened a new pull request #5906:
URL: https://github.com/apache/incubator-tvm/pull/5906
Thanks for contributing to TVM! Please refer to guideline
https://tvm.apache.org/docs/contribute/ for useful information and tips. After
the pull request is submitted, please request code re
tmoreau89 commented on pull request #5842:
URL: https://github.com/apache/incubator-tvm/pull/5842#issuecomment-648435167
Thanks @zhanghaohit I can generate the new VTA design on Pynq to verify that
it still passes on the new ISA. I will let you know if this passes.
Installation documentati
tmoreau89 commented on pull request #5858:
URL: https://github.com/apache/incubator-tvm/pull/5858#issuecomment-648434603
@badenh Please let us know if you can fix the install.rst to pass the CI
tests
This is an automated mes
icemelon9 opened a new pull request #5905:
URL: https://github.com/apache/incubator-tvm/pull/5905
Thanks for contributing to TVM! Please refer to guideline
https://tvm.apache.org/docs/contribute/ for useful information and tips. After
the pull request is submitted, please request code re
icemelon9 opened a new pull request #5904:
URL: https://github.com/apache/incubator-tvm/pull/5904
* [Relay][Fix] Fix alter op layout when calling a global var
* add test case
Thanks for contributing to TVM! Please refer to guideline
https://tvm.apache.org/docs/contribute/ fo
icemelon9 opened a new pull request #5903:
URL: https://github.com/apache/incubator-tvm/pull/5903
* Fix lambda lift
* clean up
* lint
* fix
* remove unused import
Thanks for contributing to TVM! Please refer to guideline
https://tvm.apache.org/docs/contr
zhiics commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r444506282
##
File path: python/tvm/relay/op/memory/memory.py
##
@@ -104,6 +104,21 @@ def shape_func(func, inputs, outputs, dependent=False):
"""
retu
tqchen merged pull request #5899:
URL: https://github.com/apache/incubator-tvm/pull/5899
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 g
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 81e8cfc add a few gradients (#5899)
81e8
mbrookhart commented on a change in pull request #5826:
URL: https://github.com/apache/incubator-tvm/pull/5826#discussion_r444505122
##
File path: python/tvm/relay/op/dyn/transform.py
##
@@ -0,0 +1,74 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more c
icemelon9 commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r444501324
##
File path: python/tvm/relay/op/memory/memory.py
##
@@ -104,6 +104,21 @@ def shape_func(func, inputs, outputs, dependent=False):
"""
r
zhiics commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r93947
##
File path: python/tvm/relay/op/memory/memory.py
##
@@ -104,6 +104,21 @@ def shape_func(func, inputs, outputs, dependent=False):
"""
retu
icemelon9 commented on pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#issuecomment-648395868
I feel we should treat ops like `vm.shape_of` as a dialect for vm. Therefore
the conversion from relay ops to vm dialect ops should not be inside the
manifest memory pas
icemelon9 commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r80395
##
File path: python/tvm/relay/op/memory/memory.py
##
@@ -104,6 +104,21 @@ def shape_func(func, inputs, outputs, dependent=False):
"""
r
tqchen commented on pull request #5899:
URL: https://github.com/apache/incubator-tvm/pull/5899#issuecomment-648382066
cc @jroesch @MarisaKirisame
This is an automated message from the Apache Git Service.
To respond to the me
electriclilies commented on a change in pull request #5826:
URL: https://github.com/apache/incubator-tvm/pull/5826#discussion_r67091
##
File path: src/relay/op/dyn/tensor/transform.cc
##
@@ -0,0 +1,136 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+
electriclilies commented on a change in pull request #5826:
URL: https://github.com/apache/incubator-tvm/pull/5826#discussion_r65147
##
File path: python/tvm/relay/op/dyn/transform.py
##
@@ -0,0 +1,74 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or mo
electriclilies commented on a change in pull request #5826:
URL: https://github.com/apache/incubator-tvm/pull/5826#discussion_r65147
##
File path: python/tvm/relay/op/dyn/transform.py
##
@@ -0,0 +1,74 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or mo
masahi commented on pull request #5897:
URL: https://github.com/apache/incubator-tvm/pull/5897#issuecomment-648374339
@siju-samuel please take a look
This is an automated message from the Apache Git Service.
To respond to th
ymwangg opened a new pull request #5902:
URL: https://github.com/apache/incubator-tvm/pull/5902
The following code will fail when using CUDA 11 and causes test cases of
`nvptx` target to fail. This PR fixed this issue.
```
tvm.contrib.nvcc.find_libdevice_path('7.0')
ValueError: Tra
This is an automated email from the ASF dual-hosted git repository.
jroesch pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from ef9bf7d Fix the python intrin rule (#5895)
add aa84ee2 Rust Refactor Stage 4: Rewrite Rust graph runti
jroesch merged pull request #5830:
URL: https://github.com/apache/incubator-tvm/pull/5830
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
mbrookhart commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r26108
##
File path: tests/python/relay/test_vm_serialization.py
##
@@ -332,6 +284,20 @@ def test_mobilenet():
run_network(mod, params)
+def te
zhiics commented on a change in pull request #5901:
URL: https://github.com/apache/incubator-tvm/pull/5901#discussion_r25698
##
File path: src/runtime/vm/vm.cc
##
@@ -834,18 +836,35 @@ inline ObjectRef VirtualMachine::ReadRegister(Index r)
const {
return frames_.back().
icemelon9 opened a new pull request #5901:
URL: https://github.com/apache/incubator-tvm/pull/5901
Some empirical performance improvement to Relay VM, including
- Make alignment in `AllocStorage` constant
- Use CopyTo instead of NDArray.CopyTo to avoid unnecessary memcpy
-
kevinthesun commented on pull request #5844:
URL: https://github.com/apache/incubator-tvm/pull/5844#issuecomment-648325370
Can we also add a TODO to topi cuda nms?
This is an automated message from the Apache Git Service.
To
mbrookhart opened a new pull request #5900:
URL: https://github.com/apache/incubator-tvm/pull/5900
@tqchen
These ops are already registered and tested through python, this simply adds
them to the C++ API. I don't see tests for the existing instrinsic unary ops
outside of python, do
zhiics commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r444355425
##
File path: tests/python/relay/test_vm_serialization.py
##
@@ -332,6 +284,20 @@ def test_mobilenet():
run_network(mod, params)
+def test_v
mbrookhart commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r444338874
##
File path: tests/python/relay/test_vm_serialization.py
##
@@ -332,6 +284,20 @@ def test_mobilenet():
run_network(mod, params)
+def te
tqchen merged pull request #5894:
URL: https://github.com/apache/incubator-tvm/pull/5894
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 g
This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a commit to branch v0.6
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
The following commit(s) were added to refs/heads/v0.6 by this push:
new 06352b7 [BACKPORT-0.6][Bugfix][VTA] Fixed a
tqchen commented on pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#issuecomment-648262706
Please split the rewrite simplify into its own PR and add unittestcases for
the newly added rules. It would be great if we can write down a few lines of
proof sketches, to
mbrookhart commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r444338874
##
File path: tests/python/relay/test_vm_serialization.py
##
@@ -332,6 +284,20 @@ def test_mobilenet():
run_network(mod, params)
+def te
zhiics commented on a change in pull request #5855:
URL: https://github.com/apache/incubator-tvm/pull/5855#discussion_r444325734
##
File path: src/relay/op/memory/memory.cc
##
@@ -423,5 +427,24 @@ RELAY_REGISTER_OP("memory.shape_func")
return {topi
tqchen commented on pull request #5892:
URL: https://github.com/apache/incubator-tvm/pull/5892#issuecomment-648248759
@jroesch is already working on some of the RUST FFI access. We can also do
that through python or other language binding and only use the rust runtime to
execute in wasm.
mbrookhart commented on a change in pull request #5826:
URL: https://github.com/apache/incubator-tvm/pull/5826#discussion_r444322275
##
File path: python/tvm/relay/op/dyn/transform.py
##
@@ -0,0 +1,74 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more c
jcf94 commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444299222
##
File path: src/arith/rewrite_simplify.cc
##
@@ -865,6 +900,28 @@ PrimExpr RewriteSimplifier::Impl::VisitExpr_(const
FloorModNode* op) {
TVM_T
jcf94 commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444299222
##
File path: src/arith/rewrite_simplify.cc
##
@@ -865,6 +900,28 @@ PrimExpr RewriteSimplifier::Impl::VisitExpr_(const
FloorModNode* op) {
TVM_T
merrymercy commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444283997
##
File path: src/arith/rewrite_simplify.cc
##
@@ -133,6 +133,13 @@ PrimExpr RewriteSimplifier::Impl::VisitExpr_(const
AddNode* op) {
TVM_T
merrymercy commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444283997
##
File path: src/arith/rewrite_simplify.cc
##
@@ -133,6 +133,13 @@ PrimExpr RewriteSimplifier::Impl::VisitExpr_(const
AddNode* op) {
TVM_T
zhiics merged pull request #5895:
URL: https://github.com/apache/incubator-tvm/pull/5895
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 g
zhiics merged pull request #5885:
URL: https://github.com/apache/incubator-tvm/pull/5885
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 g
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 8108f59 [Relay]Allow every runtime module to handle constants (#5885)
add ef9bf7d Fix the python intrin r
sergei-grechanik commented on a change in pull request #5618:
URL: https://github.com/apache/incubator-tvm/pull/5618#discussion_r444310282
##
File path: python/tvm/testing.py
##
@@ -188,5 +189,96 @@ def assert_prim_expr_equal(lhs, rhs):
raise ValueError("{} and {} are
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 9b745c6 remove fatal (#5888)
add 8108f59 [Relay]Allow every runtime module to handle constants (#5885)
N
zhiics commented on pull request #5885:
URL: https://github.com/apache/incubator-tvm/pull/5885#issuecomment-648235707
thanks @comaniac @mbaret @junrushao1994
This is an automated message from the Apache Git Service.
To respo
zhiics commented on pull request #5885:
URL: https://github.com/apache/incubator-tvm/pull/5885#issuecomment-648235147
@mbaret this should be handled later by changing the way to initialize graph
runtime in different PRs.
Thi
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 b94e8b7 [RFC] Improve quantized convolution performance for armv8
architectures (#5754)
add 9b745c6 remo
zhiics merged pull request #5888:
URL: https://github.com/apache/incubator-tvm/pull/5888
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 g
merrymercy commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444299601
##
File path: tests/python/unittest/test_te_schedule_gpu_advanced.py
##
@@ -0,0 +1,195 @@
+# Licensed to the Apache Software Foundation (ASF) und
merrymercy commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444299601
##
File path: tests/python/unittest/test_te_schedule_gpu_advanced.py
##
@@ -0,0 +1,195 @@
+# Licensed to the Apache Software Foundation (ASF) und
jcf94 commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444299222
##
File path: src/arith/rewrite_simplify.cc
##
@@ -865,6 +900,28 @@ PrimExpr RewriteSimplifier::Impl::VisitExpr_(const
FloorModNode* op) {
TVM_T
merrymercy removed a comment on pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#issuecomment-648223645
There are very simple unit tests for `LegalizeInvalidAttach`
```python
import tvm
from tvm import te
A = te.compute((10, 10), lambda i, j: 1.0
merrymercy commented on pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#issuecomment-648223645
There are very simple unit tests for `LegalizeInvalidAttach`
```python
import tvm
from tvm import te
A = te.compute((10, 10), lambda i, j: 1.0, name='
merrymercy commented on a change in pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#discussion_r444294661
##
File path: src/te/schedule/schedule_dataflow_rewrite.cc
##
@@ -614,10 +614,74 @@ void InjectInline(ScheduleNode* sch) {
}
}
+void Legali
t-vi commented on pull request #5899:
URL: https://github.com/apache/incubator-tvm/pull/5899#issuecomment-648179336
@siju-samuel @kevinthesun you have recent commits in the gradients
With the tests, it should be relatively low-risk.
--
t-vi opened a new pull request #5899:
URL: https://github.com/apache/incubator-tvm/pull/5899
I don't think #2562 is up to date, but hey...
This is an automated message from the Apache Git Service.
To respond to the message, p
mbaret commented on pull request #5885:
URL: https://github.com/apache/incubator-tvm/pull/5885#issuecomment-648164080
Given some modules don't use this mechanism, should we find a way to
suppress this warning?
https://github.com/apache/incubator-tvm/blob/b94e8b7290c5ced98728e730634ec737
jcf94 commented on pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#issuecomment-648123643
Seems the extra simplify rule caused a UT error, I'll try to fix that.
This is an automated message from the
jcf94 edited a comment on pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898#issuecomment-648123643
Seems the extra simplify rule caused an UT error, I'll try to fix that.
This is an automated message
FrozenGene commented on pull request #5753:
URL: https://github.com/apache/incubator-tvm/pull/5753#issuecomment-648105817
@tqchen I think you could start to review the core part now and wish to
listen your feedback. All the functionality (except the new `graph_runtime` api
signature) of `g
jcf94 opened a new pull request #5898:
URL: https://github.com/apache/incubator-tvm/pull/5898
This pr is part of #5883 , consists of several small bug fix & improvement.
> To be filled...
This is an automated messa
t-vi commented on pull request #5887:
URL: https://github.com/apache/incubator-tvm/pull/5887#issuecomment-648023393
#5897 has the fix.
This is an automated message from the Apache Git Service.
To respond to the message, pleas
t-vi opened a new pull request #5897:
URL: https://github.com/apache/incubator-tvm/pull/5897
This happens e.g. in shared input/output embeddings in BERT
or siamese networks.
Thank you @siju-samuel for reporting (in the post-merge discussion in #5887).
t-vi edited a comment on pull request #5887:
URL: https://github.com/apache/incubator-tvm/pull/5887#issuecomment-648010250
I found the problem and will send a fix.
The old code included the embedding weight twice (which arguably is a bug in
itself and needs fixing) and the new code dedup
t-vi commented on pull request #5887:
URL: https://github.com/apache/incubator-tvm/pull/5887#issuecomment-648010250
I found the problem and will send a fix.
The old code included the embedding weight twice (which arguably is a bug in
itself and needs fixing) and the new code deduplicated
mbaret closed pull request #5884:
URL: https://github.com/apache/incubator-tvm/pull/5884
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 g
mbaret commented on pull request #5884:
URL: https://github.com/apache/incubator-tvm/pull/5884#issuecomment-648001235
Superseded by https://github.com/apache/incubator-tvm/pull/5885
This is an automated message from the Apach
siju-samuel commented on pull request #5887:
URL: https://github.com/apache/incubator-tvm/pull/5887#issuecomment-647993442
Output mismatch between pytorch and tvm.
Afer this PR
```
File "pytorch_pretrained_bert_uncased.py", line 112, in
tvm.testing.assert_allclose(torc
t-vi commented on pull request #5887:
URL: https://github.com/apache/incubator-tvm/pull/5887#issuecomment-647989008
What is the error you are seeing?
This is an automated message from the Apache Git Service.
To respond to
siju-samuel commented on pull request #5887:
URL: https://github.com/apache/incubator-tvm/pull/5887#issuecomment-647953471
@t-vi Thanks for the PR.
I think this PR has slight impact in some scenarios.
You can confirm by running the below script before and after your change.
[
100 matches
Mail list logo