FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365092221
##
File path: python/tvm/contrib/cc.py
##
@@ -51,10 +51,41 @@ def create_shared(output,
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365091179
##
File path: python/tvm/contrib/cc.py
##
@@ -15,7 +15,7 @@
# specific language governing
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365091130
##
File path: python/tvm/contrib/cc.py
##
@@ -51,10 +51,41 @@ def create_shared(output,
else:
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365071448
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365069861
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365069930
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365069096
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365067929
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365067929
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
This is an automated email from the ASF dual-hosted git repository.
kevinthesun pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from d6a23cf [REFACTOR][IR] tvm::Expr -> PrimExpr(Primitive Expr) (#4669)
add 06ce76b download
kevinthesun commented on issue #4671: [AutoTVM] Download fallback schedule file
if it does not exist
URL: https://github.com/apache/incubator-tvm/pull/4671#issuecomment-572867160
Thanks @zxy844288792 @comaniac
This is an
kevinthesun merged pull request #4671: [AutoTVM] Download fallback schedule
file if it does not exist
URL: https://github.com/apache/incubator-tvm/pull/4671
This is an automated message from the Apache Git Service.
To
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365066925
##
File path: python/tvm/contrib/ndk.py
##
@@ -64,5 +64,28 @@ def create_shared(output,
msg
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365066395
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365066395
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365066395
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365065825
##
File path: include/tvm/codegen.h
##
@@ -59,6 +59,21 @@ runtime::Module Build(const Array&
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365065438
##
File path: src/codegen/llvm/llvm_module.cc
##
@@ -62,6 +63,11 @@ class LLVMModuleNode final :
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365065474
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365062069
##
File path: src/codegen/llvm/codegen_blob.cc
##
@@ -0,0 +1,163 @@
+/*
+ * Licensed to the
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365062086
##
File path: src/codegen/llvm/codegen_blob.h
##
@@ -0,0 +1,51 @@
+/*
+ * Licensed to the Apache
FrozenGene commented on a change in pull request #4657: [CodeGen] Generate blob
use LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r365060564
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
optima2005 commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r365048590
##
File path: python/tvm/relay/frontend/onnx.py
##
@@ -223,37 +223,64 @@ class Conv(OnnxOpConverter):
optima2005 commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r365054371
##
File path: tests/python/frontend/onnx/test_forward.py
##
@@ -1732,22 +1732,34 @@ def test_or():
optima2005 commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r365054591
##
File path: tests/python/frontend/onnx/test_forward.py
##
@@ -1761,18 +1773,95 @@ def verify_conv(x_shape,
optima2005 commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r365054153
##
File path: tests/python/frontend/onnx/test_forward.py
##
@@ -1732,22 +1732,34 @@ def test_or():
optima2005 commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r365055937
##
File path: topi/python/topi/cuda/conv1d.py
##
@@ -0,0 +1,308 @@
+# Licensed to the Apache Software Foundation
tqchen commented on issue #4643: [REFACTOR] Replace TensorObj and TensorValue
with NDArray
URL: https://github.com/apache/incubator-tvm/pull/4643#issuecomment-572852296
still need to wait for @jroesch 's approval
This is an
zhiics commented on issue #4643: [REFACTOR] Replace TensorObj and TensorValue
with NDArray
URL: https://github.com/apache/incubator-tvm/pull/4643#issuecomment-572842661
@tqchen Done.
This is an automated message from the
optima2005 commented on issue #4659: [CONV3D] remove incorrect schedule for
conv3d_ndhwc
URL: https://github.com/apache/incubator-tvm/pull/4659#issuecomment-572833998
@masahi Yes. it would report error. Currently at least in TF frontend, the
NHWC for conv2d, NDHWC for conv3d are forced to
tqchen commented on issue #4673: [REFACTOR][IR] Initialize Unified IR Expr Data
Structure
URL: https://github.com/apache/incubator-tvm/pull/4673#issuecomment-572825641
cc @jroesch @wweic @icemelon9 @yzhliu @merrymercy
This
tqchen opened a new pull request #4673: [REFACTOR][IR] Initialize Unified IR
Expr Data Structure
URL: https://github.com/apache/incubator-tvm/pull/4673
This PR moves a few base types from relay and low-level Expr into the ir
sub-folder.
These classes will serve as a common type system
yongfeng-nv commented on a change in pull request #4651: Tensor Expression
Debug Display (TEDD)
URL: https://github.com/apache/incubator-tvm/pull/4651#discussion_r365031269
##
File path: python/tvm/contrib/tedd.py
##
@@ -0,0 +1,506 @@
+# Licensed to the Apache Software
yzhliu removed a comment on issue #4486: [Arith] Bound for Shape Variables
URL: https://github.com/apache/incubator-tvm/pull/4486#issuecomment-572806468
@tqchen How about name it "IndexVar" so as to be more precise? term "shape"
could be considered as multi-dimensional
comaniac commented on a change in pull request #4671: [AutoTVM] Download
fallback schedule file if it does not exist
URL: https://github.com/apache/incubator-tvm/pull/4671#discussion_r365016101
##
File path: python/tvm/autotvm/tophub.py
##
@@ -214,7 +214,12 @@ def
anijain2305 commented on issue #4664: [Docs] Convert Layout pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#issuecomment-572809935
Thanks @jwfromm @yzhliu @trevor-m
I have massaged section 3 to be more precise and instructive. Please take a
second look if you can.
zxy844288792 commented on a change in pull request #4671: [AutoTVM] Download
fallback schedule file if it does not exist
URL: https://github.com/apache/incubator-tvm/pull/4671#discussion_r365014847
##
File path: python/tvm/autotvm/tophub.py
##
@@ -208,6 +208,10 @@ def
jwfromm commented on issue #4672: [Relay][Frontend][Onnx] Added pool
autopadding and simplified parsers.
URL: https://github.com/apache/incubator-tvm/pull/4672#issuecomment-572808906
@masahi, can you take a look at this? It's very related to other recent PRs.
jwfromm opened a new pull request #4672: [Relay][Frontend][Onnx] Added pool
autopadding and simplified parsers.
URL: https://github.com/apache/incubator-tvm/pull/4672
This PR adds support for the auto padding operator to Onnx. It also
simplifies all types of pooling converters to a single
yzhliu commented on issue #4486: [Arith] Bound for Shape Variables
URL: https://github.com/apache/incubator-tvm/pull/4486#issuecomment-572806468
@tqchen How about name it "IndexVar" so as to be more precise? term "shape"
could be considered as multi-dimensional
comaniac commented on a change in pull request #4671: [AutoTVM] Download
fallback schedule file if it does not exist
URL: https://github.com/apache/incubator-tvm/pull/4671#discussion_r365012071
##
File path: python/tvm/autotvm/tophub.py
##
@@ -208,6 +208,10 @@ def
tqchen commented on issue #4643: [REFACTOR] Replace TensorObj and TensorValue
with NDArray
URL: https://github.com/apache/incubator-tvm/pull/4643#issuecomment-572805537
@zhiics please do a rebase
This is an automated message
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 619c63c Use int for endch to fix portability issues regarding
signed/unsigned char (#4668)
add d6a23cf
tqchen merged pull request #4669: [REFACTOR][IR] tvm::Expr ->
PrimExpr(Primitive Expr)
URL: https://github.com/apache/incubator-tvm/pull/4669
This is an automated message from the Apache Git Service.
To respond to the
zxy844288792 opened a new pull request #4671: [AutoTVM] Download fallback
schedule file if it does not exist
URL: https://github.com/apache/incubator-tvm/pull/4671
Autotvm will download the default schedule file from tophub if user does not
specify a schedule file. Howver, if user
yongwww commented on issue #4417: [TOPI][RELAY][OP] add op crop_and_resize
URL: https://github.com/apache/incubator-tvm/pull/4417#issuecomment-572801902
@kevinthesun @zhiics @jwfromm @anijain2305 fixed comments and ci. Please
take another look
yongfeng-nv commented on a change in pull request #4651: Tensor Expression
Debug Display (TEDD)
URL: https://github.com/apache/incubator-tvm/pull/4651#discussion_r365005585
##
File path: python/tvm/contrib/tedd.py
##
@@ -0,0 +1,506 @@
+# Licensed to the Apache Software
comaniac commented on issue #4602: [Docs] Bring Your Own Codegen Guide -- Part 1
URL: https://github.com/apache/incubator-tvm/pull/4602#issuecomment-572784413
@u99127 have you got a chance to review it?
This is an automated
yzhliu commented on a change in pull request #4664: [Docs] Convert Layout pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364907721
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation (ASF)
yzhliu commented on a change in pull request #4664: [Docs] Convert Layout pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364984440
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation (ASF)
yzhliu commented on a change in pull request #4664: [Docs] Convert Layout pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364907959
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation (ASF)
yzhliu commented on a change in pull request #4664: [Docs] Convert Layout pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364905522
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation (ASF)
yzhliu commented on a change in pull request #4664: [Docs] Convert Layout pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364906219
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation (ASF)
jwfromm commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r364983601
##
File path: src/relay/op/nn/convolution.cc
##
@@ -52,6 +50,67 @@ Array > ConvInferCorrectLayout(
masahi commented on a change in pull request #4667: [Tutorial] Deploy Quantized
Model on CUDA
URL: https://github.com/apache/incubator-tvm/pull/4667#discussion_r364973536
##
File path: tutorials/frontend/deploy_quantized.py
##
@@ -0,0 +1,154 @@
+# Licensed to the Apache
masahi commented on a change in pull request #4667: [Tutorial] Deploy Quantized
Model on CUDA
URL: https://github.com/apache/incubator-tvm/pull/4667#discussion_r364966229
##
File path: tutorials/frontend/deploy_quantized.py
##
@@ -0,0 +1,154 @@
+# Licensed to the Apache
zhiics commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364923528
##
File path: src/codegen/llvm/codegen_blob.cc
##
@@ -0,0 +1,163 @@
+/*
+ * Licensed to the Apache
zhiics commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364951790
##
File path: src/codegen/llvm/codegen_blob.h
##
@@ -0,0 +1,51 @@
+/*
+ * Licensed to the Apache
zhiics commented on issue #4643: [REFACTOR] Replace TensorObj and TensorValue
with NDArray
URL: https://github.com/apache/incubator-tvm/pull/4643#issuecomment-572724424
@jroesch can you take another look? Should we merge?
trevor-m commented on a change in pull request #4664: [Docs] Convert Layout
pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364924367
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation
trevor-m commented on a change in pull request #4664: [Docs] Convert Layout
pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364924367
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation
trevor-m commented on a change in pull request #4664: [Docs] Convert Layout
pass.
URL: https://github.com/apache/incubator-tvm/pull/4664#discussion_r364924367
##
File path: docs/dev/convert_layout.rst
##
@@ -0,0 +1,192 @@
+.. Licensed to the Apache Software Foundation
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364912655
##
File path: python/tvm/contrib/ndk.py
##
@@ -64,5 +64,28 @@ def create_shared(output,
msg
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364913418
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364913289
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364911422
##
File path: include/tvm/codegen.h
##
@@ -59,6 +59,21 @@ runtime::Module Build(const Array& funcs,
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364914242
##
File path: src/codegen/llvm/llvm_module.cc
##
@@ -62,6 +63,11 @@ class LLVMModuleNode final :
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364913154
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen commented on a change in pull request #4657: [CodeGen] Generate blob use
LLVM directly
URL: https://github.com/apache/incubator-tvm/pull/4657#discussion_r364911682
##
File path: python/tvm/contrib/cc.py
##
@@ -51,9 +51,32 @@ def create_shared(output,
else:
tqchen closed issue #4660: Unicode, Base64 errors using SaveJSON, LoadJSON on
constants with aarch64 only (works on x86_64)
URL: https://github.com/apache/incubator-tvm/issues/4660
This is an automated message from the
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 7a30e5f [Relay][Frontend][TFlite] Add parses support for unary
elemwise ops (#4634)
add 619c63c Use
gussmith23 opened a new issue #4670: [Relay] Bias missing in final layer of
Mobilenet?
URL: https://github.com/apache/incubator-tvm/issues/4670
https://github.com/apache/incubator-tvm/blob/3f43bee01ea50d21ef93f618be3540bcf5231cd7/python/tvm/relay/testing/mobilenet.py#L123
Should
tqchen opened a new pull request #4669: [REFACTOR][IR] tvm::Expr ->
PrimExpr(Primitive Expr)
URL: https://github.com/apache/incubator-tvm/pull/4669
As part of unified IR, we will need to unify relay::Expr
and the current tvm::Expr under the same base type.
From the techinical
tqchen commented on issue #4669: [REFACTOR][IR] tvm::Expr -> PrimExpr(Primitive
Expr)
URL: https://github.com/apache/incubator-tvm/pull/4669#issuecomment-572704116
cc @jroesch @yzhliu @merrymercy @ZihengJiang @wweic
This is
tqchen commented on issue #4634: [Relay][Frontend][TFlite] Add parses support
for unary elemwise ops
URL: https://github.com/apache/incubator-tvm/pull/4634#issuecomment-572698103
Thanks @FrozenGene @anijain2305 @inadob
This
tqchen merged pull request #4634: [Relay][Frontend][TFlite] Add parses support
for unary elemwise ops
URL: https://github.com/apache/incubator-tvm/pull/4634
This is an automated message from the Apache Git Service.
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 78243e9 [REFACTOR] relay::Module Def -> TypeDef (#4665)
add 7a30e5f [Relay][Frontend][TFlite] Add
trevor-m opened a new pull request #4668: Fix Base64OutStream portability issue
URL: https://github.com/apache/incubator-tvm/pull/4668
Fixes #4660
On ARM, char is unsigned by default while on x86 it is signed by default. On
ARM, this causes the logic in `Base64OutStream::Finish` to
anijain2305 commented on a change in pull request #4667: [Tutorial] Deploy
Quantized Model on CUDA
URL: https://github.com/apache/incubator-tvm/pull/4667#discussion_r364868831
##
File path: tutorials/frontend/deploy_quantized.py
##
@@ -0,0 +1,154 @@
+# Licensed to the
anijain2305 commented on a change in pull request #4667: [Tutorial] Deploy
Quantized Model on CUDA
URL: https://github.com/apache/incubator-tvm/pull/4667#discussion_r364867350
##
File path: tutorials/frontend/deploy_quantized.py
##
@@ -0,0 +1,154 @@
+# Licensed to the
anijain2305 commented on a change in pull request #4667: [Tutorial] Deploy
Quantized Model on CUDA
URL: https://github.com/apache/incubator-tvm/pull/4667#discussion_r364867851
##
File path: tutorials/frontend/deploy_quantized.py
##
@@ -0,0 +1,154 @@
+# Licensed to the
tqchen commented on issue #4665: [REFACTOR] relay::Module Def -> TypeDef
URL: https://github.com/apache/incubator-tvm/pull/4665#issuecomment-572661132
Thanks @wweic
This is an automated message from the Apache Git Service.
tqchen merged pull request #4665: [REFACTOR] relay::Module Def -> TypeDef
URL: https://github.com/apache/incubator-tvm/pull/4665
This is an automated message from the Apache Git Service.
To respond to the message, please log
icemelon9 closed pull request #4659: [CONV3D] remove incorrect schedule for
conv3d_ndhwc
URL: https://github.com/apache/incubator-tvm/pull/4659
This is an automated message from the Apache Git Service.
To respond to the
icemelon9 commented on issue #4659: [CONV3D] remove incorrect schedule for
conv3d_ndhwc
URL: https://github.com/apache/incubator-tvm/pull/4659#issuecomment-572656770
Never mind. I'll fix the conv3d issue in #4644 since I'm changing the TOPI.
vinx13 opened a new pull request #4667: [Tutorial] Deploy Quantized Model on
CUDA
URL: https://github.com/apache/incubator-tvm/pull/4667
This tutorial demonstrates how to import a model using Relay frontend, run
quantization and calibration passes, and perform quantized inference.
ref
masahi commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r364752946
##
File path: src/relay/op/nn/convolution.cc
##
@@ -52,6 +50,67 @@ Array > ConvInferCorrectLayout(
masahi commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r364752270
##
File path: topi/python/topi/generic/nn.py
##
@@ -34,6 +34,42 @@ def _default_schedule(outs, auto_inline):
return
masahi commented on a change in pull request #4639: [Relay/Topi][Op] Conv1D
URL: https://github.com/apache/incubator-tvm/pull/4639#discussion_r364752378
##
File path: topi/python/topi/generic/nn.py
##
@@ -34,6 +34,42 @@ def _default_schedule(outs, auto_inline):
return
masahi commented on issue #4663: [Relay/Topi][Op] 1D Pooling
URL: https://github.com/apache/incubator-tvm/pull/4663#issuecomment-572477342
thanks @jwfromm
This is an automated message from the Apache Git Service.
To respond
masahi merged pull request #4663: [Relay/Topi][Op] 1D Pooling
URL: https://github.com/apache/incubator-tvm/pull/4663
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
This is an automated email from the ASF dual-hosted git repository.
masahi pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git.
from baae28b [Autotvm] Use VM compile to extract autotvm tasks (#4328)
add 8a98a2e [Relay/Topi][Op] 1D
Maximilianxu commented on issue #4061: nvcc fatal : Value 'sm_75' is not
defined for option 'gpu-architecture'
URL: https://github.com/apache/incubator-tvm/issues/4061#issuecomment-572453353
@jikechao Finally, finally, I solved this issue. Although I have no idea why
this issue
wyc-ruiker opened a new pull request #4666: [Relay][Frontend][TFLite] Add
constant input support for elemwise ops
URL: https://github.com/apache/incubator-tvm/pull/4666
TFLite elementwise ops (like `ADD`, `MUL`, etc.) only support `rhs_expr` is
a constant value. But `lhs_expr` can also be
masahi commented on a change in pull request #4497: [WIP] [Relay] Add a PyTorch
to Relay Parser
URL: https://github.com/apache/incubator-tvm/pull/4497#discussion_r364599526
##
File path: python/tvm/relay/frontend/pytorch.py
##
@@ -0,0 +1,1104 @@
+# Licensed to the Apache
95 matches
Mail list logo