roastduck commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420568421
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} els
zhiics commented on pull request #5501:
URL: https://github.com/apache/incubator-tvm/pull/5501#issuecomment-624451593
yeah, these errors should be reproducible locally.
This is an automated message from the Apache Git Service
zhiics commented on a change in pull request #5517:
URL: https://github.com/apache/incubator-tvm/pull/5517#discussion_r420549748
##
File path: include/tvm/runtime/packed_func.h
##
@@ -45,6 +45,15 @@
#define TVM_RUNTIME_HEADER_ONLY 0
#endif
+// Always inline macro only use i
tqchen commented on pull request #5517:
URL: https://github.com/apache/incubator-tvm/pull/5517#issuecomment-62593
cc @yzhliu @ZihengJiang @zhiics
This is an automated message from the Apache Git Service.
To respond to th
yongfeng-nv commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420543138
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} e
lixiaoquan commented on a change in pull request #5429:
URL: https://github.com/apache/incubator-tvm/pull/5429#discussion_r420541474
##
File path: tests/python/relay/test_any.py
##
@@ -138,23 +138,35 @@ def test_any_concat():
result = ex.evaluate()(x_np, y_np)
lixiaoquan commented on a change in pull request #5429:
URL: https://github.com/apache/incubator-tvm/pull/5429#discussion_r420541295
##
File path: src/relay/transforms/pattern_util.h
##
@@ -270,7 +270,7 @@ inline Constant MakeConstantScalar(DataType dtype, T value)
{
*/
tem
masahi commented on pull request #5521:
URL: https://github.com/apache/incubator-tvm/pull/5521#issuecomment-624428430
Thanks @mbrookhart
This is an automated message from the Apache Git Service.
To respond to the message, pl
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 16cb571 [TFLITE]Nit: Function names made consitent (#5515)
add 7eb2451 fix prelu importer and add tests
tqchen commented on pull request #5507:
URL: https://github.com/apache/incubator-tvm/pull/5507#issuecomment-624416758
I am a bit confused by why the bits field is necessary, would it suffice to
just check the dtype of each tensor and compute the maximum size correctly?
---
siju-samuel commented on pull request #5507:
URL: https://github.com/apache/incubator-tvm/pull/5507#issuecomment-624415991
@tqchen could you please review. Thanks.
This is an automated message from the Apache Git Service.
To
siju-samuel commented on pull request #5502:
URL: https://github.com/apache/incubator-tvm/pull/5502#issuecomment-624415529
@masahi @FrozenGene @kazum Please help to review. TIA
This is an automated message from the Apache Git
siju-samuel commented on pull request #5486:
URL: https://github.com/apache/incubator-tvm/pull/5486#issuecomment-624415336
@FrozenGene could you please help to merge this PR. TIA
This is an automated message from the Apache G
tqchen opened a new pull request #5522:
URL: https://github.com/apache/incubator-tvm/pull/5522
cc @areusch @tmoreau89 @ZihengJiang
This is an automated message from the Apache Git Service.
To respond to the message, please l
tqchen commented on pull request #5501:
URL: https://github.com/apache/incubator-tvm/pull/5501#issuecomment-624408573
@cchung100m likely you will need to fix the compilation error due to type
mismatch.
Also for a node field update we will need to update the json compact
function to
tqchen commented on issue #5490:
URL: https://github.com/apache/incubator-tvm/issues/5490#issuecomment-624408382
NOTE, as in the intiial example, we will need to update the converter(via
_update_from_std_str) to keep backward compatibility of the nodes
roastduck commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420490285
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} els
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420481027
##
File path: docs/dev/convert_layout.rst
##
@@ -116,16 +117,21 @@ These steps happen for each operator in sequence, where
ConvertLayout pass k
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420479959
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalizat
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420479959
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalizat
comaniac commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420477165
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalization
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420475487
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalizat
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420475487
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalizat
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 16cb571 [TFLITE]Nit: Function names made
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 7cbc0ca [PATCH] [ring_buffer.h] Improve
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 32a094c [QNN] Support CallNode inputs in qnn.concatenate (#5360)
add 4c9724d [RPC][BUGFIX][BACKPORT-0.6
mbrookhart opened a new pull request #5521:
URL: https://github.com/apache/incubator-tvm/pull/5521
@masahi Thank you!
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHu
comaniac commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420445637
##
File path: docs/dev/convert_layout.rst
##
@@ -116,16 +117,21 @@ These steps happen for each operator in sequence, where
ConvertLayout pass keep
mbrookhart opened a new pull request #5520:
URL: https://github.com/apache/incubator-tvm/pull/5520
LRN only supports [4D
tensors](https://github.com/apache/incubator-tvm/blob/32a094cc22b5464605918b6c1dcc42e6b0925733/topi/include/topi/nn/local_response_norm.h#L59),
so we shouldn't propogate
comaniac commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420443258
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalization
comaniac commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420442683
##
File path: docs/dev/convert_layout.rst
##
@@ -218,24 +224,49 @@ Second example is for a lightly-layout sensitive operator
- batch normalization
lhutton1 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420434693
##
File path: tests/python/relay/test_pass_convert_op_layout.py
##
@@ -625,3 +685,5 @@ def expected():
test_qnn_conv_requantize_convert_layout
lhutton1 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420434153
##
File path: docs/dev/convert_layout.rst
##
@@ -116,16 +117,21 @@ These steps happen for each operator in sequence, where
ConvertLayout pass keep
u99127 opened a new pull request #5519:
URL: https://github.com/apache/incubator-tvm/pull/5519
This is a draft PR and only for discussion but not for merging as is.
These are a couple of commits that show a proof of concept about how we
could restructure and improve the tflite fronte
anijain2305 commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420426034
##
File path: docs/dev/convert_layout.rst
##
@@ -116,16 +117,21 @@ These steps happen for each operator in sequence, where
ConvertLayout pass k
anijain2305 commented on pull request #5360:
URL: https://github.com/apache/incubator-tvm/pull/5360#issuecomment-624320687
Thanks @mbaret @zhiics This is merged
This is an automated message from the Apache Git Service.
To res
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 70a5902 [RPC] Call sync in remote cpu to gpu copies (#5512)
add 32a094c [QNN] Support CallNode inp
merrymercy edited a comment on pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#issuecomment-624317263
tag @FrozenGene who also met this problem
This is an automated message from the Apache Git Servi
merrymercy commented on pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#issuecomment-624317263
@FrozenGene also net this problem
This is an automated message from the Apache Git Service.
To respond
u99127 opened a new pull request #5518:
URL: https://github.com/apache/incubator-tvm/pull/5518
bytes_available refers to the number of bytes used in the ring buffer. At
the same time, fix a couple of typos.
@tqchen
--
u99127 commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420413713
##
File path: src/support/ring_buffer.h
##
@@ -63,17 +67,28 @@ class RingBuffer {
size_t ncopy = head_ptr_ + bytes_available_ - old_size;
tqchen opened a new pull request #5517:
URL: https://github.com/apache/incubator-tvm/pull/5517
- Add static assert to warn about unsupported type deduction.
- Always inline template expansions for PackedFunc calls.
Thi
u99127 commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420394839
##
File path: src/support/ring_buffer.h
##
@@ -63,19 +67,27 @@ class RingBuffer {
size_t ncopy = head_ptr_ + bytes_available_ - old_size;
tqchen commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420393992
##
File path: tests/python/unittest/test_runtime_rpc.py
##
@@ -102,6 +102,14 @@ def remote_array_func(y):
fremote(r_cpu)
+def test_rpc_large
tqchen commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420393740
##
File path: src/support/ring_buffer.h
##
@@ -63,17 +67,28 @@ class RingBuffer {
size_t ncopy = head_ptr_ + bytes_available_ - old_size;
u99127 commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420387251
##
File path: src/support/ring_buffer.h
##
@@ -63,17 +67,28 @@ class RingBuffer {
size_t ncopy = head_ptr_ + bytes_available_ - old_size;
u99127 commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420387251
##
File path: src/support/ring_buffer.h
##
@@ -63,17 +67,28 @@ class RingBuffer {
size_t ncopy = head_ptr_ + bytes_available_ - old_size;
areusch commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420387111
##
File path: tests/python/unittest/test_runtime_rpc.py
##
@@ -102,6 +102,14 @@ def remote_array_func(y):
fremote(r_cpu)
+def test_rpc_larg
tqchen commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420382680
##
File path: tests/python/unittest/test_runtime_rpc.py
##
@@ -102,6 +102,14 @@ def remote_array_func(y):
fremote(r_cpu)
+def test_rpc_large
tqchen commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420382680
##
File path: tests/python/unittest/test_runtime_rpc.py
##
@@ -102,6 +102,14 @@ def remote_array_func(y):
fremote(r_cpu)
+def test_rpc_large
tqchen commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420382680
##
File path: tests/python/unittest/test_runtime_rpc.py
##
@@ -102,6 +102,14 @@ def remote_array_func(y):
fremote(r_cpu)
+def test_rpc_large
areusch commented on a change in pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516#discussion_r420379452
##
File path: tests/python/unittest/test_runtime_rpc.py
##
@@ -102,6 +102,14 @@ def remote_array_func(y):
fremote(r_cpu)
+def test_rpc_larg
tqchen commented on issue #5514:
URL: https://github.com/apache/incubator-tvm/issues/5514#issuecomment-624259864
Please see if https://github.com/apache/incubator-tvm/pull/5516 fixes the
problem
This is an automated message
tqchen opened a new pull request #5516:
URL: https://github.com/apache/incubator-tvm/pull/5516
cc @merrymercy @giuseros
This is an automated message from the Apache Git Service.
To respond to the message, please log on to Gi
tqchen commented on issue #5514:
URL: https://github.com/apache/incubator-tvm/issues/5514#issuecomment-624250487
Looking into the diff you point out, perhaps the most relevant one would be
the change on the ring buffer
This
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 70a5902 [RPC] Call sync in remote cpu to
u99127 commented on pull request #5510:
URL: https://github.com/apache/incubator-tvm/pull/5510#issuecomment-624204486
>
>
> > Minor nits, A better commit message would be to refer to handling batch
matmul in fully connected op rather than what it stands.
> > Isn't there a need fo
u99127 commented on pull request #5515:
URL: https://github.com/apache/incubator-tvm/pull/5515#issuecomment-624203498
>
>
> Some nits in TFLITE. Inconsistencies in the function names in tflite is
removed.
>
> @u99127 Could you please review this. Thanks in advance.
s/c
siju-samuel opened a new pull request #5515:
URL: https://github.com/apache/incubator-tvm/pull/5515
Some nits in TFLITE. Inconsistencies in the function names in tflite is
removed.
@u99127 Could you please review this. Thanks in advance.
siju-samuel commented on a change in pull request #5486:
URL: https://github.com/apache/incubator-tvm/pull/5486#discussion_r420289173
##
File path: python/tvm/relay/frontend/tflite.py
##
@@ -109,16 +109,17 @@ def __init__(self, model, subgraph, exp_tab):
'PAD': sel
giuseros opened a new issue #5514:
URL: https://github.com/apache/incubator-tvm/issues/5514
Hi all,
I am running TVM from an Ubuntu 16.04 machine and I have the tracker running
on the same machine.
An aarch64 machine is connected to the tracker.
When running from the mast
siju-samuel commented on a change in pull request #5479:
URL: https://github.com/apache/incubator-tvm/pull/5479#discussion_r420274518
##
File path: python/tvm/relay/testing/tf.py
##
@@ -184,10 +184,15 @@ def get_workload_official(model_url, model_sub_path):
dir_path = os.p
tqchen edited a comment on pull request #5512:
URL: https://github.com/apache/incubator-tvm/pull/5512#issuecomment-624189332
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
tqchen commented on pull request #5512:
URL: https://github.com/apache/incubator-tvm/pull/5512#issuecomment-624189332
@areusch We only do it for CopyFromRemote and CopyToRemote. In both cases,
we(user or rpc server) allocate temp meomory from the arena, which will become
invalid after the
areusch commented on pull request #5512:
URL: https://github.com/apache/incubator-tvm/pull/5512#issuecomment-624186951
what's the justification for doing this just over RPC sessions?
This is an automated message from the Apac
comaniac commented on a change in pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#discussion_r420265752
##
File path: docs/dev/convert_layout.rst
##
@@ -116,16 +117,21 @@ These steps happen for each operator in sequence, where
ConvertLayout pass keep
u99127 commented on a change in pull request #5486:
URL: https://github.com/apache/incubator-tvm/pull/5486#discussion_r420269543
##
File path: python/tvm/relay/frontend/tflite.py
##
@@ -109,16 +109,17 @@ def __init__(self, model, subgraph, exp_tab):
'PAD': self.con
anijain2305 commented on a change in pull request #4805:
URL: https://github.com/apache/incubator-tvm/pull/4805#discussion_r420254831
##
File path: python/tvm/relay/frontend/tflite.py
##
@@ -463,6 +484,114 @@ def convert_relu(self, op):
return out
+def convert_
yongfeng-nv commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420255081
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} e
anijain2305 edited a comment on pull request #4805:
URL: https://github.com/apache/incubator-tvm/pull/4805#issuecomment-624168410
Sorry for the delay. @inadob Can you take a look at this PR -
https://github.com/apache/incubator-tvm/pull/5479
I have added support for fused quantized a
anijain2305 commented on pull request #4805:
URL: https://github.com/apache/incubator-tvm/pull/4805#issuecomment-624168410
Sorry for the delay. @inadob Can you take a look at this PR -
https://github.com/apache/incubator-tvm/pull/5479
I have added support for quantized activations th
anijain2305 commented on pull request #5479:
URL: https://github.com/apache/incubator-tvm/pull/5479#issuecomment-624167303
@FrozenGene Please review
@mbaret Please take a second look
This is an automated message from the A
siju-samuel opened a new pull request #5513:
URL: https://github.com/apache/incubator-tvm/pull/5513
reverse/flip added support for axes similar to mxnet/numpy.
@FrozenGene Please help me to review this PR.
This is
tqchen opened a new pull request #5512:
URL: https://github.com/apache/incubator-tvm/pull/5512
cc @tmoreau89 @areusch @ZihengJiang
This is an automated message from the Apache Git Service.
To respond to the message, please l
roastduck commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420230912
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} els
cchung100m opened a new pull request #5511:
URL: https://github.com/apache/incubator-tvm/pull/5511
Hi @FrozenGene @kevinthesun @comaniac
I am interested in issue #5215, however, I am not familiar with `TOPI` and
`backend`. I would appreciate that if you can guide me to complete this
roastduck commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420190981
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} els
roastduck commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420187744
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} els
roastduck commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r420187744
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} els
FrozenGene commented on pull request #5485:
URL: https://github.com/apache/incubator-tvm/pull/5485#issuecomment-624109424
> @FrozenGene
>
> We tested the layout as you suggested, and the results are listed below.
>
>
>
>
>
> kernel:3x3x64x64 feature maps:56x56
wsl-inspur commented on pull request #5485:
URL: https://github.com/apache/incubator-tvm/pull/5485#issuecomment-624103994
@FrozenGene
We tested the layout as you suggested, and the results are listed below.
kernel:3x3x64x64 feature maps:56x56x64:
batch | kernel (alpha, alp
ANSHUMAN87 commented on pull request #5510:
URL: https://github.com/apache/incubator-tvm/pull/5510#issuecomment-624102545
Gentle ping @FrozenGene !
This is an automated message from the Apache Git Service.
To respond to the m
yongfeng-nv commented on a change in pull request #5382:
URL: https://github.com/apache/incubator-tvm/pull/5382#discussion_r419559292
##
File path: src/te/operation/op_util.cc
##
@@ -164,9 +164,21 @@ MakeLoopNest(const Stage& stage,
value_map[iv] = dom->min;
} e
Menooker commented on a change in pull request #5357:
URL: https://github.com/apache/incubator-tvm/pull/5357#discussion_r420068669
##
File path: topi/python/topi/x86/conv2d_alter_op.py
##
@@ -31,6 +32,9 @@
logger = logging.getLogger('topi')
+_NCHWc_matcher = re.compile("^N
Menooker commented on a change in pull request #5357:
URL: https://github.com/apache/incubator-tvm/pull/5357#discussion_r420067574
##
File path: python/tvm/relay/op/strategy/x86.py
##
@@ -128,6 +135,15 @@ def conv2d_strategy_cpu(attrs, inputs, out_type, target):
Menooker commented on a change in pull request #5357:
URL: https://github.com/apache/incubator-tvm/pull/5357#discussion_r420067349
##
File path: python/tvm/relay/op/strategy/x86.py
##
@@ -96,6 +99,12 @@ def conv2d_strategy_cpu(attrs, inputs, out_type, target):
lhutton1 commented on pull request #5422:
URL: https://github.com/apache/incubator-tvm/pull/5422#issuecomment-623970891
There seems to be an unrelated issue with the depthwise convolution check
when using the default NHWC kernel layout: `if is_depthwise_conv2d(data.shape,
attrs['data_layou
87 matches
Mail list logo