[GitHub] [incubator-tvm] abergeron opened a new issue #5704: tests/python/integration/test_ewise.py::test_add fail on macOS with Metal
abergeron opened a new issue #5704:
URL: https://github.com/apache/incubator-tvm/issues/5704
To reproduce, build on macOS X with metal enabled (I use min-version=10.11,
if that makes a difference), install the python modules and run the test with
pytest.
The error seems to complain about the "float44" dtype which I think doesn't
exist.
```
[23:06:17]
/Users/anakha/miniconda/conda-bld/tvm-libs_1590891729956/work/src/runtime/metal/metal_device_api.mm:131:
Intializing Metal device 0, name=AMD Radeon R9 M290
___ test_add
___
def test_add():
def run(dtype):
# graph
n = te.size_var('n')
A = te.placeholder((n,), name='A', dtype=dtype)
B = te.placeholder((n,), name='B', dtype=dtype)
bias = te.var("bias", dtype=dtype)
scale = te.var("scale", dtype=dtype)
C = te.compute(A.shape, lambda *i: A(*i) + B(*i), name='C')
# schedule
s = te.create_schedule(C.op)
# create iter var and assign them tags.
num_thread = 16
bx, x = s[C].split(C.op.axis[0], factor=num_thread*4)
tx, x = s[C].split(x, nparts=num_thread)
_, x = s[C].split(x, factor=4)
s[C].bind(bx, te.thread_axis("blockIdx.x"))
s[C].bind(tx, te.thread_axis("threadIdx.x"))
s[C].vectorize(x)
# one line to build the function.
def check_device(device):
ctx = tvm.context(device, 0)
if not ctx.exist:
print("skip because %s is not enabled.." % device)
return
fadd = tvm.build(s, [A, B, C],
device,
name="myadd")
# launch the kernel.
n = 1024
a = tvm.nd.array((np.random.uniform(size=n) *
256).astype(A.dtype), ctx)
b = tvm.nd.array((np.random.uniform(size=n) *
256).astype(B.dtype), ctx)
c = tvm.nd.array(np.zeros(n, dtype=C.dtype), ctx)
ftimer = fadd.time_evaluator(fadd.entry_name, ctx, number=1)
tcost = ftimer(a, b, c).mean
tvm.testing.assert_allclose(
c.asnumpy(), a.asnumpy() + b.asnumpy(), rtol=1e-6)
check_device("opencl")
check_device("cuda")
if dtype == "float32":
check_device("metal")
check_device("vulkan")
> run("float32")
tests/python/integration/test_ewise.py:254:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _
tests/python/integration/test_ewise.py:251: in run
check_device("metal")
tests/python/integration/test_ewise.py:244: in check_device
tcost = ftimer(a, b, c).mean
../_test_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_p/lib/python3.7/site-packages/tvm/runtime/module.py:215:
in evaluator
blob = feval(*args)
tvm/_ffi/_cython/./packed_func.pxi:321: in
tvm._ffi._cy3.core.PackedFuncBase.__call__
???
tvm/_ffi/_cython/./packed_func.pxi:256: in tvm._ffi._cy3.core.FuncCall
???
tvm/_ffi/_cython/./packed_func.pxi:245: in tvm._ffi._cy3.core.FuncCall3
???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _
> ???
E tvm._ffi.base.TVMError: Traceback (most recent call last):
E [bt] (3) 4 libtvm.dylib0x00011d8f4f48
TVMFuncCall + 72
E [bt] (2) 3 libtvm.dylib0x00011d93c687
std::__1::__function::__func, void (tvm::runtime::TVMArgs,
tvm::runtime::TVMRetValue*)>::operator()(tvm::runtime::TVMArgs&&,
tvm::runtime::TVMRetValue*&&) + 359
E [bt] (1) 2 libtvm.dylib0x00011d900103
std::__1::__function::__func
const&)::$_0, std::__1::allocator const&)::$_0>, void
(tvm::runtime::TVMArgs,
tvm::runtime::TVMRetValue*)>::operator()(tvm::runtime::TVMArgs&&,
tvm::runtime::TVMRetValue*&&) + 259
E [bt] (0) 1 libtvm.dylib0x00011ceb84bf
dmlc::LogMessageFatal::~LogMessageFatal() + 111
E [bt] (7) 8 ??? 0x001a26946585
0x0 + 112316409221
E [bt] (6) 7 libtvm.dylib0x00011d8f4be4
TVMBackendGetFuncFromEnv + 164
E [bt] (5) 6 libtvm.dylib0x00011d901d77
tvm::runtime::ModuleNode::GetFuncFromEnv(std::__1::basic_string, std::__1::allocator > const&) + 231
E [bt] (4) 5
[GitHub] [incubator-tvm-vta] liangfu commented on pull request #8: [Hardware][Xilinx] explicitly specify acc dep distance to avoid hidden pitfall
liangfu commented on pull request #8: URL: https://github.com/apache/incubator-tvm-vta/pull/8#issuecomment-636420830 I would agree with > derive the value from the VTA target (i.e. FPGA type) in pkg_config , and avoid requiring a user to specify the `ACC_DEP_DISTANCE` parameter. 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] comaniac opened a new pull request #5703: [PatternLang] Simplify Pattern API Implementations
comaniac opened a new pull request #5703: URL: https://github.com/apache/incubator-tvm/pull/5703 For each pattern node, we defined `XXPattern` Python class for the FFI connection. We also define a user-friendly API such as `is_op` to have better user experience. Since `is_XX` APIs are just wrappers to the corresponding `XXPattern` class, this PR simplifies their implementations by turning the APIs to function aliases. In addition, this PR also makes the following changes: 1. Add `is_constant`, `is_tuple`, and `is_tuple_get_item` APIs. 2. Rename `is_input` to `is_var` to against `is_constant`. @mbrookhart I'm not quite sure if the second change makes sense to you or not so I intentionally separate it to the second commit. If you prefer `is_input`, I could just simply revert it. Also cc @masahi 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] kevinthesun commented on pull request #5684: [AutoTVM][TOPI] Fix bifrost spatial packing conv2d auto tune
kevinthesun commented on pull request #5684: URL: https://github.com/apache/incubator-tvm/pull/5684#issuecomment-636407068 Also can you fix winograd kernel replacement? 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] kevinthesun commented on a change in pull request #5684: [AutoTVM][TOPI] Fix bifrost spatial packing conv2d auto tune
kevinthesun commented on a change in pull request #5684: URL: https://github.com/apache/incubator-tvm/pull/5684#discussion_r432900758 ## File path: topi/python/topi/arm_cpu/conv2d_spatial_pack.py ## @@ -267,9 +270,13 @@ def conv2d_spatial_pack_nhwc(cfg, data, kernel, strides, padding, dilation, out_ data_vec = te.compute(dvshape, lambda n, oho, owo, ohi, owi, ic: data_pad[n][oho*OHI*HSTR+ohi][owo*OWI*WSTR+owi][ic], name='data_vec') -kernel_vec = te.compute(kvshape, lambda oco, kh, kw, ic, oci: \ -kernel[kh][kw][ic][oco*OCI+oci], -name='kernel_vec') + +if autotvm.GLOBAL_SCOPE.in_tuning: Review comment: Change schedule for arm_cpu conv2d as well? 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[incubator-tvm] branch master updated (c55ed37 -> 55aefc2)
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 c55ed37 [REFACTOR][RELAY] Replace build_config with PassContext (#5698) add 55aefc2 [PYTORCH]floor_divide support for squeezenet (#5702) No new revisions were added by this update. Summary of changes: python/tvm/relay/frontend/pytorch.py | 1 + 1 file changed, 1 insertion(+)
[GitHub] [incubator-tvm] masahi merged pull request #5702: [PYTORCH]floor_divide support for squeezenet
masahi merged pull request #5702: URL: https://github.com/apache/incubator-tvm/pull/5702 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] yongwww commented on a change in pull request #4312: [TOPI][Relay][OP] Dynamic NMS and strided_slice
yongwww commented on a change in pull request #4312:
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r432886819
##
File path: python/tvm/relay/op/_transform.py
##
@@ -99,8 +99,80 @@ def _arange_shape_func(start, stop, step):
@_reg.register_shape_func("arange", True)
def arange_shape_func(attrs, inputs, _):
+"""
+Shape func for arange
+"""
return [_arange_shape_func(*inputs)]
+@script
+def _strided_slice_shape_func_input_data(data, begin, end, strides,
+ slice_mode):
+ndim = len(data.shape)
+out = output_tensor((ndim,), "int64")
+for i in const_range(ndim):
+cbegin = 0
+cend = data.shape[i]
+cstride = 1
+if strides.shape[0] > i:
+cstride = strides[i]
+if begin.shape[0] > i:
+cbegin = begin[i]
+if end.shape[0] <= i:
+cend = data.shape[i]
+elif slice_mode != 0:
+if end[i] < 0:
+cend = data.shape[i]
+elif cstride < 0:
+cend = cbegin - end[i]
+else:
+cend = cbegin + end[i]
+else:
+cend = end[i]
+assert cstride != 0, "Strides can't be zero."
+out[i] = int64(ceil_div((int64(cend) - int64(cbegin)), int64(cstride)))
+return out
+
+@script
+def _strided_slice_shape_func_input_shape(data_shape, begin, end, strides,
slice_mode):
+ndim = data_shape.shape[0]
+assert ndim == 2, "not correct"
+out = output_tensor((ndim,), "int64")
+for i in const_range(ndim):
+cbegin = int64(0)
+cend = int64(data_shape[i])
+cstride = int64(1)
+if len(strides) > i:
+cstride = int64(strides[i])
+if len(begin) > i:
+cbegin = int64(begin[i])
+if len(end) <= i:
+cend = int64(data_shape[i])
+elif slice_mode != 0:
+if end[i] < 0:
+cend = int64(data_shape[i])
+elif cstride < 0:
Review comment:
yes, updated
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 the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
[GitHub] [incubator-tvm] yongwww commented on a change in pull request #4312: [TOPI][Relay][OP] Dynamic NMS and strided_slice
yongwww commented on a change in pull request #4312:
URL: https://github.com/apache/incubator-tvm/pull/4312#discussion_r432886691
##
File path: python/tvm/relay/frontend/tensorflow.py
##
@@ -614,6 +614,52 @@ def _impl(inputs, attr, params, mod):
return out
return _impl
+def _nms():
+def _impl(inputs, attr, params, mod):
+# Get parameter values
+max_output_size =
int(np.atleast_1d(inputs[2].data.asnumpy().astype("int64"))[0])
+iou_threshold = np.atleast_1d(inputs[3].data.asnumpy())[0]
+# score_threshold was introduced from V3
+score_threshold = np.atleast_1d(inputs[4].data.asnumpy())[0] if
len(inputs) > 4 else 0.0
+
+# Generate data with shape (1, num_anchors, 5)
+scores = AttrCvt(op_name="expand_dims",
+ ignores=['T_threshold'],
+ extras={'axis': -1, 'num_newaxis': 1})([inputs[1]],
attr)
+data = get_relay_op('concatenate')([scores, inputs[0]], -1)
+data = get_relay_op('expand_dims')(data, 0, 1)
+
+# reason why using get_valid_counts is for inference performance
+ct, data, indices = get_relay_op('get_valid_counts')(data,
+
score_threshold=score_threshold,
+ id_index=-1,
+ score_index=0)
+# TensorFlow NMS doesn't have parameter top_k
+top_k = -1
+# TF doesn't have class id for nms input
+score_index = 0
+nms_ret = get_relay_op('non_max_suppression')(data=data,
+ valid_count=ct,
+ indices=indices,
+
max_output_size=max_output_size,
+
iou_threshold=iou_threshold,
+ force_suppress=True,
+ top_k=top_k,
+ coord_start=1,
+ score_index=score_index,
+ id_index=-1,
+ return_indices=True,
+ invalid_to_bottom=False)
+
+# squeeze it, TF NMS is not batched
+end = get_relay_op("squeeze")(nms_ret[1], axis=[1])
+data_slice = get_relay_op("squeeze")(nms_ret[0], axis=[0])
+
+# slice to get the dynamic result
+ret = get_relay_op("strided_slice")(data_slice, _expr.const([0]), end,
_expr.const([1]))
+return ret
+return _impl
Review comment:
Updated. Also used slice_mode for tf `Slice`
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 the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
[GitHub] [incubator-tvm] majiang31312 commented on issue #5686: [vulkan] Assertion in tir/transforms/lower_thread_allreduce.cc", line 157 TVMError: Check failed: v:
majiang31312 commented on issue #5686: URL: https://github.com/apache/incubator-tvm/issues/5686#issuecomment-636342592 I'm new to TVM, but I will have a try :) Thanks! @tqchen 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] siju-samuel opened a new pull request #5702: [PYTORCH]floor_divide support for squeezenet
siju-samuel opened a new pull request #5702: URL: https://github.com/apache/incubator-tvm/pull/5702 `aten::floor_divide` support. https://github.com/apache/incubator-tvm/issues/5133#issuecomment-636330705 Testcase, im not able to simulate floor_divide. @masahi Please help to review this. 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] power0341 commented on issue #5133: [Torch] A list of missing op conversion in need of help
power0341 commented on issue #5133: URL: https://github.com/apache/incubator-tvm/issues/5133#issuecomment-636336567 @siju-samuel thanks a lot, it works, good job 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] siju-samuel commented on issue #5133: [Torch] A list of missing op conversion in need of help
siju-samuel commented on issue #5133:
URL: https://github.com/apache/incubator-tvm/issues/5133#issuecomment-636334501
@power0341
Please apply the below patch and let me know.
```
diff --git a/python/tvm/relay/frontend/pytorch.py
b/python/tvm/relay/frontend/pytorch.py
index f68affd82..f2e24a128 100644
--- a/python/tvm/relay/frontend/pytorch.py
+++ b/python/tvm/relay/frontend/pytorch.py
@@ -1760,6 +1760,7 @@ def _get_convert_map(prelude):
"aten::arange" : _arange(),
"aten::div" : _elemwise("divide"),
"aten::div_": _elemwise("divide"),
+"aten::floor_divide": _elemwise("floor_divide"),
"aten::addcdiv" : _addcdiv(),
"aten::addcmul" : _addcmul(),
"aten::ones": _ones(),
```
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 the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
[GitHub] [incubator-tvm] power0341 commented on issue #5133: [Torch] A list of missing op conversion in need of help
power0341 commented on issue #5133: URL: https://github.com/apache/incubator-tvm/issues/5133#issuecomment-636330705 > NotImplementedError: The following operators are not implemented: ['aten::floor_divide'] can we have this guy 'aten::floor_divide' as well? it's required by shufflenetV2 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] srkreddy1238 commented on a change in pull request #5695: fix small bug about dense_grad
srkreddy1238 commented on a change in pull request #5695: URL: https://github.com/apache/incubator-tvm/pull/5695#discussion_r432831890 ## File path: python/tvm/relay/op/_tensor_grad.py ## @@ -472,8 +472,8 @@ def bias_add_grad(orig, grad): def dense_grad(orig, grad): """Returns [grad' @ weight, data @ grad']""" data, weight = orig.args -return [collapse_sum_like(transpose(grad) * weight, data), -collapse_sum_like(data * transpose(grad), weight)] +return [collapse_sum_like(_nn.dense(grad, transpose(weight)), data), Review comment: In above example of data (5, 4) and weight (3, 4) implies a dense with 4 inputs and yielding 3 outputs each for 5 units. Hence 5 here is the batch when we apply to a network. We very well support units/batches. Can you share details on the error you got while units is added as arg ? 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] handar423 commented on a change in pull request #5695: fix small bug about dense_grad
handar423 commented on a change in pull request #5695: URL: https://github.com/apache/incubator-tvm/pull/5695#discussion_r432826485 ## File path: python/tvm/relay/op/_tensor_grad.py ## @@ -472,8 +472,8 @@ def bias_add_grad(orig, grad): def dense_grad(orig, grad): """Returns [grad' @ weight, data @ grad']""" data, weight = orig.args -return [collapse_sum_like(transpose(grad) * weight, data), -collapse_sum_like(data * transpose(grad), weight)] +return [collapse_sum_like(_nn.dense(grad, transpose(weight)), data), Review comment: Thank you for your response! After correcting it, I found that both x86 and CUDA only support dense without batching, I tried testing with arm-cpu(mobile) but came into [python/tvm/relay/op/strategy/x86.py](https://github.com/apache/incubator-tvm/blob/master/python/tvm/relay/op/strategy/x86.py) and failed as well, so could you please tell me how to run dense with batching? Thanks again! 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] FrozenGene commented on issue #5215: [AutoTVM] AutoTVM incorrect measurement
FrozenGene commented on issue #5215: URL: https://github.com/apache/incubator-tvm/issues/5215#issuecomment-636300190 > @FrozenGene In #5200 we discussed another source of autotvm inaccurate measurement due to empty input tensor. Do we have a timeline to fix that? @kevinthesun FYI. We are rebasing our code based on the laster master. After completing it, this fix will be brought in. 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[incubator-tvm] branch rust-tvm created (now a44a379)
This is an automated email from the ASF dual-hosted git repository. jroesch pushed a change to branch rust-tvm in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git. at a44a379 Refactor anyhow out of the rt layer This branch includes the following new commits: new 4cf2dbc Add tvm crate new a44a379 Refactor anyhow out of the rt layer The 2 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference.
[incubator-tvm] 01/02: Add tvm crate
This is an automated email from the ASF dual-hosted git repository.
jroesch pushed a commit to branch rust-tvm
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
commit 4cf2dbcf9b058e4d10f332a3fe0385d9387929fc
Author: Jared Roesch
AuthorDate: Thu May 28 02:08:14 2020 -0700
Add tvm crate
---
rust/tvm/.gitignore | 7 +
rust/tvm/.travis.yml | 22 +++
rust/tvm/Cargo.toml | 45 +
rust/tvm/README.md | 235 +++
rust/tvm/examples/resnet/Cargo.toml | 29
rust/tvm/examples/resnet/README.md | 45 +
rust/tvm/examples/resnet/build.rs| 42 +
rust/tvm/examples/resnet/src/build_resnet.py | 134 +++
rust/tvm/examples/resnet/src/main.rs | 160 ++
rust/tvm/src/ir/array.rs | 70
rust/tvm/src/ir/mod.rs | 17 ++
rust/tvm/src/ir/relay/mod.rs | 232 ++
rust/tvm/src/lib.rs | 60 +++
rust/tvm/src/runtime/mod.rs | 1 +
rust/tvm/src/transform.rs| 41 +
rust/tvm/tests/basics/.gitignore | 7 +
rust/tvm/tests/basics/Cargo.toml | 32
rust/tvm/tests/basics/build.rs | 46 ++
rust/tvm/tests/basics/src/main.rs| 55 +++
rust/tvm/tests/basics/src/tvm_add.py | 50 ++
rust/tvm/tests/callback/Cargo.toml | 26 +++
rust/tvm/tests/callback/src/bin/array.rs | 72
rust/tvm/tests/callback/src/bin/error.rs | 56 +++
rust/tvm/tests/callback/src/bin/float.rs | 50 ++
rust/tvm/tests/callback/src/bin/int.rs | 49 ++
rust/tvm/tests/callback/src/bin/string.rs| 54 ++
rust/tvm/tests/test_ir.rs| 37 +
27 files changed, 1674 insertions(+)
diff --git a/rust/tvm/.gitignore b/rust/tvm/.gitignore
new file mode 100644
index 000..2430329
--- /dev/null
+++ b/rust/tvm/.gitignore
@@ -0,0 +1,7 @@
+target
+**/*.rs.bk
+Cargo.lock
+/tests/basics/add_*
+/examples/resnet/deploy_*
+/examples/resnet/*.png
+/examples/resnet/synset.*
diff --git a/rust/tvm/.travis.yml b/rust/tvm/.travis.yml
new file mode 100644
index 000..e963b7c
--- /dev/null
+++ b/rust/tvm/.travis.yml
@@ -0,0 +1,22 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements. See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership. The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License. You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied. See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+language: rust
+rust:
+ - nightly
+matrix:
+ fast_finish: true
diff --git a/rust/tvm/Cargo.toml b/rust/tvm/Cargo.toml
new file mode 100644
index 000..4cbb619
--- /dev/null
+++ b/rust/tvm/Cargo.toml
@@ -0,0 +1,45 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements. See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership. The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License. You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied. See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+[package]
+name = "tvm"
+version = "0.1.0"
+license = "Apache-2.0"
+description = "Rust frontend support for TVM"
+repository = "https://github.com/apache/incubator-tvm;
+homepage = "https://github.com/apache/incubator-tvm;
+readme = "README.md"
+keywords = ["rust", "tvm"]
+categories = ["api-bindings", "science"]
+authors = ["TVM Contributors"]
+edition = "2018"
+
+[dependencies]
+thiserror = "^1.0"
+anyhow = "^1.0"
+lazy_static = "1.1"
+ndarray = "0.12"
+num-traits = "0.2"
+tvm-rt = { version = "0.1", path = "../tvm-rt/" }
+tvm-sys = { version = "0.1", path = "../tvm-sys/" }
+tvm-macros = { version = "*", path = "../macros/" }
+paste = "0.1"
+mashup = "0.1"
+once_cell = "^1.3.1"
+
+[features]
+blas = ["ndarray/blas"]
diff --git
[incubator-tvm] 02/02: Refactor anyhow out of the rt layer
This is an automated email from the ASF dual-hosted git repository.
jroesch pushed a commit to branch rust-tvm
in repository https://gitbox.apache.org/repos/asf/incubator-tvm.git
commit a44a379bb3b3f4fab505dce3520eeb97f230ac23
Author: Jared Roesch
AuthorDate: Sat May 30 01:07:46 2020 -0700
Refactor anyhow out of the rt layer
---
rust/Cargo.toml | 3 +-
rust/macros/src/object.rs| 8 ++---
rust/tvm-rt/src/errors.rs| 36
rust/tvm-rt/src/function.rs | 66 +---
rust/tvm-rt/src/lib.rs | 5 +--
rust/tvm-rt/src/ndarray.rs | 66 +++-
rust/tvm-rt/src/object/mod.rs| 17 --
rust/tvm-rt/src/object/object_ptr.rs | 30 +---
rust/tvm-rt/src/to_function.rs | 37 ++--
rust/tvm-rt/src/value.rs | 5 ++-
rust/tvm/src/ir/array.rs | 55 +-
rust/tvm/src/lib.rs | 10 +-
rust/tvm/src/transform.rs| 2 +-
13 files changed, 211 insertions(+), 129 deletions(-)
diff --git a/rust/Cargo.toml b/rust/Cargo.toml
index 6d3481b..e107104 100644
--- a/rust/Cargo.toml
+++ b/rust/Cargo.toml
@@ -29,5 +29,6 @@ members = [
"frontend/tests/callback",
"frontend/examples/resnet",
"tvm-sys",
- "tvm-rt"
+ "tvm-rt",
+ "tvm",
]
diff --git a/rust/macros/src/object.rs b/rust/macros/src/object.rs
index 96a86dd..670d326 100644
--- a/rust/macros/src/object.rs
+++ b/rust/macros/src/object.rs
@@ -89,12 +89,12 @@ pub fn macro_impl(input: proc_macro::TokenStream) ->
TokenStream {
}
impl std::convert::TryFrom for #ref_id {
-type Error = ::anyhow::Error;
+type Error = tvm_rt::Error;
fn try_from(ret_val: tvm_rt::RetValue) -> Result<#ref_id,
Self::Error> {
use std::convert::TryInto;
let oref: ObjectRef = ret_val.try_into()?;
-let ptr = oref.0.ok_or(anyhow::anyhow!("null ptr"))?;
+let ptr = oref.0.ok_or(tvm_rt::Error::Null)?;
let ptr = ptr.downcast::<#payload_id>()?;
Ok(#ref_id(Some(ptr)))
}
@@ -122,7 +122,7 @@ pub fn macro_impl(input: proc_macro::TokenStream) ->
TokenStream {
}
impl<'a> std::convert::TryFrom> for #ref_id {
-type Error = anyhow::Error;
+type Error = tvm_rt::Error;
fn try_from(arg_value: tvm_rt::ArgValue<'a>) -> Result<#ref_id,
Self::Error> {
use std::convert::TryInto;
@@ -132,7 +132,7 @@ pub fn macro_impl(input: proc_macro::TokenStream) ->
TokenStream {
}
impl<'a> std::convert::TryFrom<_rt::ArgValue<'a>> for #ref_id {
-type Error = anyhow::Error;
+type Error = tvm_rt::Error;
fn try_from(arg_value: _rt::ArgValue<'a>) -> Result<#ref_id,
Self::Error> {
use std::convert::TryInto;
diff --git a/rust/tvm-rt/src/errors.rs b/rust/tvm-rt/src/errors.rs
index 77dbba7..41e873f 100644
--- a/rust/tvm-rt/src/errors.rs
+++ b/rust/tvm-rt/src/errors.rs
@@ -17,13 +17,10 @@
* under the License.
*/
+use crate::DataType;
use thiserror::Error;
#[derive(Debug, Error)]
-#[error("Cannot convert from an empty array.")]
-pub struct EmptyArrayError;
-
-#[derive(Debug, Error)]
#[error("Handle `{name}` is null.")]
pub struct NullHandleError {
pub name: String,
@@ -41,5 +38,32 @@ pub struct TypeMismatchError {
}
#[derive(Debug, Error)]
-#[error("Missing NDArray shape.")]
-pub struct MissingShapeError;
+pub enum NDArrayError {
+#[error("Missing NDArray shape.")]
+MissingShape,
+#[error("Cannot convert from an empty array.")]
+EmptyArray,
+#[error("Invalid datatype when attempting to convert ndarray.")]
+InvalidDatatype(#[from] tvm_sys::datatype::ParseDataTypeError),
+#[error("a shape error occurred in the Rust ndarray library")]
+ShapeError(#[from] ndarray::ShapeError),
+#[error("Expected type `{expected}` but found `{actual}`")]
+DataTypeMismatch { expected: DataType, actual: DataType }
+}
+
+#[derive(Debug, Error)]
+pub enum Error {
+#[error("{0}")]
+Downcast(#[from] tvm_sys::errors::ValueDowncastError),
+#[error("raw pointer passed across boundary was null")]
+Null,
+}
+
+impl Error {
+pub fn downcast(actual_type: String, expected_type: &'static str) -> Error
{
+Self::Downcast(tvm_sys::errors::ValueDowncastError {
+actual_type,
+expected_type,
+})
+}
+}
diff --git a/rust/tvm-rt/src/function.rs b/rust/tvm-rt/src/function.rs
index 2a5f446..17f5f6e 100644
--- a/rust/tvm-rt/src/function.rs
+++ b/rust/tvm-rt/src/function.rs
@@ -33,12 +33,14 @@ use std::{
ptr, slice, str,
sync::Mutex,
};
-
+use std::convert::{TryFrom};
use anyhow::Result;
use lazy_static::lazy_static;
pub use
[GitHub] [incubator-tvm] FrozenGene commented on issue #5038: [RFC] Module based Model Runtime Interface
FrozenGene commented on issue #5038: URL: https://github.com/apache/incubator-tvm/issues/5038#issuecomment-636294461 > @FrozenGene can we follow up on this? Hi, @tqchen I will start to work on it from next Monday! Sorry for working on it lately because of other things. 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [incubator-tvm] cchung100m commented on a change in pull request #5684: [AutoTVM][TOPI] Fix bifrost spatial packing conv2d auto tune
cchung100m commented on a change in pull request #5684: URL: https://github.com/apache/incubator-tvm/pull/5684#discussion_r432817217 ## File path: topi/python/topi/bifrost/conv2d.py ## @@ -142,13 +142,14 @@ def _schedule_spatial_pack(cfg, s, output, conv, data_vec, kernel_vec): s[data_vec].unroll(vw) if isinstance(kernel_vec.op, tvm.te.ComputeOp) and kernel_vec.name == 'kernel_vec': +co, ci, kh, kw, vc = s[kernel_vec].op.axis Review comment: Hi @kevinthesun Thanks for the review, I move the placeholder replacement to compute. 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 the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
