You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2020/01/13 07:47:48 UTC
[GitHub] [incubator-mxnet] Yiyan66 opened a new pull request #17280: [numpy]
add op random.exponential
Yiyan66 opened a new pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280
## Description ##
add op random.exponential
## Checklist ##
### Essentials ###
Please feel free to remove inapplicable items for your PR.
- [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) created (except PRs with tiny changes)
- [ ] Changes are complete (i.e. I finished coding on this PR)
- [ ] All changes have test coverage:
- Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
- Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
- Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
- [ ] Code is well-documented:
- For user-facing API changes, API doc string has been updated.
- For new C++ functions in header files, their functionalities and arguments are documented.
- For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
- Check the API doc at https://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
- [ ] To the best of my knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change
### Changes ###
- [ ] Feature1, tests, (and when applicable, API doc)
- [ ] Feature2, tests, (and when applicable, API doc)
## Comments ##
- If this change is a backward incompatible change, why must this change be made.
- Interesting edge cases to note here
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] xidulu commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
xidulu commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365824552
##########
File path: tests/python/unittest/test_numpy_op.py
##########
@@ -3367,6 +3367,32 @@ def hybrid_forward(self, F, x):
assert out.shape == expected_shape
+@with_seed()
+@use_np
+def test_np_exponential():
+ class TestRandomExp(HybridBlock):
+ def __init__(self, shape):
+ super(TestRandomExp, self).__init__()
+ self._shape = shape
+
+ def hybrid_forward(self, F, scale):
+ return F.np.random.exponential(scale, self._shape)
+
+ shapes = [(), (1,), (2, 3), (4, 0, 5), 6, (7, 8), None]
+ for hybridize in [False, True]:
+ for shape in shapes:
+ test_exponential = TestRandomExp(shape)
+ if hybridize:
+ test_exponential.hybridize()
+ np_out = _np.random.exponential(size = shape)
+ mx_out = test_exponential(np.array([1]))
+
+ for shape in shapes:
+ mx_out = np.random.exponential(np.array([1]), shape)
+ np_out = _np.random.exponential(np.array([1]).asnumpy(), shape)
+ assert_almost_equal(mx_out.asnumpy().shape, np_out.shape)
Review comment:
You'd better test the raise cases like this: https://github.com/apache/incubator-mxnet/blob/master/tests/python/unittest/test_numpy_op.py#L3131
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 merged pull request #17280: [numpy] add
op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 merged pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365680375
##########
File path: src/operator/numpy/random/np_exponential_op.h
##########
@@ -0,0 +1,147 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential_op.h
+ * \brief Operator for numpy sampling from exponential distribution.
+ */
+
+#ifndef MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+#define MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+
+#include <mxnet/operator_util.h>
+#include <algorithm>
+#include <string>
+#include <vector>
+#include <cmath>
+#include "../../elemwise_op_common.h"
+#include "../../mshadow_op.h"
+#include "../../mxnet_op.h"
+#include "../../operator_common.h"
+#include "../../tensor/elemwise_binary_broadcast_op.h"
+#include "./dist_common.h"
+
+namespace mxnet {
+namespace op {
+
+struct NumpyExponentialParam : public dmlc::Parameter<NumpyExponentialParam> {
+ dmlc::optional<float> scale;
+ dmlc::optional<mxnet::Tuple<int>> size;
+ DMLC_DECLARE_PARAMETER(NumpyExponentialParam) {
+ DMLC_DECLARE_FIELD(scale)
+ .set_default(dmlc::optional<float> (1.0));
+ DMLC_DECLARE_FIELD(size)
+ .set_default(dmlc::optional<mxnet::Tuple<int>>())
+ .describe("Output shape. If the given shape is, "
+ "e.g., (m, n, k), then m * n * k samples are drawn. "
+ "Default is None, in which case a single value is returned.");
+ }
+};
+
+template <typename DType>
+struct scalar_exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, float scale, float *threshold,
+ DType *out) {
+ out[i] = -scale * log(threshold[i]);
+ }
+};
+
+namespace mxnet_op {
+
+template <typename IType>
+struct check_legal_scale_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, IType *scalar, float* flag) {
+ if (scalar[i] < 0.0) {
+ flag[0] = -1.0;
+ }
+ }
+};
+
+
+template <int ndim, typename IType, typename OType>
+struct exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i,
+ const Shape<ndim> &stride,
+ const Shape<ndim> &oshape,
+ IType *scales, float* threshold, OType *out) {
+ Shape<ndim> coord = unravel(i, oshape);
+ auto idx = static_cast<index_t>(dot(coord, stride));
+ out[i] = -scales[idx] * log(threshold[i]);
+ }
+};
+
+} // namespace mxnet_op
+
+template <typename xpu>
+void NumpyExponentialForward(const nnvm::NodeAttrs &attrs,
+ const OpContext &ctx,
+ const std::vector<TBlob> &inputs,
+ const std::vector<OpReqType> &req,
+ const std::vector<TBlob> &outputs) {
+ using namespace mshadow;
+ using namespace mxnet_op;
+ const NumpyExponentialParam ¶m = nnvm::get<NumpyExponentialParam>(attrs.parsed);
+ Stream<xpu> *s = ctx.get_stream<xpu>();
+ index_t output_len = outputs[0].Size();
+ Random<xpu, float> *prnd = ctx.requested[0].get_random<xpu, float>(s);
+ Tensor<xpu, 1, float> workspace =
+ ctx.requested[1].get_space_typed<xpu, 1, float>(Shape1(output_len + 1), s);
+ Tensor<xpu, 1, float> uniform_tensor = workspace.Slice(0, output_len);
+ Tensor<xpu, 1, float> indicator_device = workspace.Slice(output_len, output_len + 1);
+ float indicator_host = 1.0;
+ float *indicator_device_ptr = indicator_device.dptr_;
+ Kernel<set_zero, xpu>::Launch(s, 1, indicator_device_ptr);
+ prnd->SampleUniform(&workspace, 0.0, 1.0);
+ if (param.scale.has_value()) {
+ CHECK_GE(param.scale.value(), 0.0) << "ValueError: expect scale >= 0";
+ MSHADOW_REAL_TYPE_SWITCH(outputs[0].type_flag_, DType, {
+ Kernel<scalar_exponential_kernel<DType>, xpu>::Launch(
+ s, outputs[0].Size(), param.scale.value(),
+ uniform_tensor.dptr_, outputs[0].dptr<DType>());
+ });
+ } else {
+ MSHADOW_TYPE_SWITCH(inputs[0].type_flag_, IType, {
+ Kernel<check_legal_scale_kernel<IType>, xpu>::Launch(
+ s, inputs[0].Size(), inputs[0].dptr<IType>(), indicator_device_ptr);
+ });
Review comment:
mis-aligned here.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365680272
##########
File path: src/operator/numpy/random/np_exponential_op.h
##########
@@ -0,0 +1,147 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential_op.h
+ * \brief Operator for numpy sampling from exponential distribution.
+ */
+
+#ifndef MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+#define MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+
+#include <mxnet/operator_util.h>
+#include <algorithm>
+#include <string>
+#include <vector>
+#include <cmath>
+#include "../../elemwise_op_common.h"
+#include "../../mshadow_op.h"
+#include "../../mxnet_op.h"
+#include "../../operator_common.h"
+#include "../../tensor/elemwise_binary_broadcast_op.h"
+#include "./dist_common.h"
+
+namespace mxnet {
+namespace op {
+
+struct NumpyExponentialParam : public dmlc::Parameter<NumpyExponentialParam> {
+ dmlc::optional<float> scale;
+ dmlc::optional<mxnet::Tuple<int>> size;
+ DMLC_DECLARE_PARAMETER(NumpyExponentialParam) {
+ DMLC_DECLARE_FIELD(scale)
+ .set_default(dmlc::optional<float> (1.0));
+ DMLC_DECLARE_FIELD(size)
+ .set_default(dmlc::optional<mxnet::Tuple<int>>())
+ .describe("Output shape. If the given shape is, "
+ "e.g., (m, n, k), then m * n * k samples are drawn. "
+ "Default is None, in which case a single value is returned.");
+ }
+};
+
+template <typename DType>
+struct scalar_exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, float scale, float *threshold,
+ DType *out) {
+ out[i] = -scale * log(threshold[i]);
+ }
+};
+
+namespace mxnet_op {
+
+template <typename IType>
+struct check_legal_scale_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, IType *scalar, float* flag) {
+ if (scalar[i] < 0.0) {
+ flag[0] = -1.0;
+ }
+ }
+};
+
+
+template <int ndim, typename IType, typename OType>
+struct exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i,
+ const Shape<ndim> &stride,
+ const Shape<ndim> &oshape,
+ IType *scales, float* threshold, OType *out) {
+ Shape<ndim> coord = unravel(i, oshape);
+ auto idx = static_cast<index_t>(dot(coord, stride));
+ out[i] = -scales[idx] * log(threshold[i]);
+ }
+};
+
+} // namespace mxnet_op
+
+template <typename xpu>
+void NumpyExponentialForward(const nnvm::NodeAttrs &attrs,
+ const OpContext &ctx,
+ const std::vector<TBlob> &inputs,
+ const std::vector<OpReqType> &req,
+ const std::vector<TBlob> &outputs) {
+ using namespace mshadow;
+ using namespace mxnet_op;
+ const NumpyExponentialParam ¶m = nnvm::get<NumpyExponentialParam>(attrs.parsed);
+ Stream<xpu> *s = ctx.get_stream<xpu>();
+ index_t output_len = outputs[0].Size();
+ Random<xpu, float> *prnd = ctx.requested[0].get_random<xpu, float>(s);
+ Tensor<xpu, 1, float> workspace =
+ ctx.requested[1].get_space_typed<xpu, 1, float>(Shape1(output_len + 1), s);
+ Tensor<xpu, 1, float> uniform_tensor = workspace.Slice(0, output_len);
+ Tensor<xpu, 1, float> indicator_device = workspace.Slice(output_len, output_len + 1);
+ float indicator_host = 1.0;
+ float *indicator_device_ptr = indicator_device.dptr_;
+ Kernel<set_zero, xpu>::Launch(s, 1, indicator_device_ptr);
+ prnd->SampleUniform(&workspace, 0.0, 1.0);
+ if (param.scale.has_value()) {
+ CHECK_GE(param.scale.value(), 0.0) << "ValueError: expect scale >= 0";
+ MSHADOW_REAL_TYPE_SWITCH(outputs[0].type_flag_, DType, {
+ Kernel<scalar_exponential_kernel<DType>, xpu>::Launch(
+ s, outputs[0].Size(), param.scale.value(),
+ uniform_tensor.dptr_, outputs[0].dptr<DType>());
+ });
+ } else {
+ MSHADOW_TYPE_SWITCH(inputs[0].type_flag_, IType, {
+ Kernel<check_legal_scale_kernel<IType>, xpu>::Launch(
+ s, inputs[0].Size(), inputs[0].dptr<IType>(), indicator_device_ptr);
+ });
+ _copy<xpu>(s, &indicator_host, indicator_device_ptr);
+ CHECK_GE(indicator_host, 0.0)
+ << "ValueError: expect scale >= 0";
Review comment:
I think this check should fit in a single line.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365683789
##########
File path: src/operator/numpy/random/np_exponential_op.h
##########
@@ -0,0 +1,147 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential_op.h
+ * \brief Operator for numpy sampling from exponential distribution.
+ */
+
+#ifndef MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+#define MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+
+#include <mxnet/operator_util.h>
+#include <algorithm>
+#include <string>
+#include <vector>
+#include <cmath>
+#include "../../elemwise_op_common.h"
+#include "../../mshadow_op.h"
+#include "../../mxnet_op.h"
+#include "../../operator_common.h"
+#include "../../tensor/elemwise_binary_broadcast_op.h"
+#include "./dist_common.h"
+
+namespace mxnet {
+namespace op {
+
+struct NumpyExponentialParam : public dmlc::Parameter<NumpyExponentialParam> {
+ dmlc::optional<float> scale;
+ dmlc::optional<mxnet::Tuple<int>> size;
+ DMLC_DECLARE_PARAMETER(NumpyExponentialParam) {
+ DMLC_DECLARE_FIELD(scale)
+ .set_default(dmlc::optional<float> (1.0));
Review comment:
`dmlc::optional<float>(1.0)`
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365680152
##########
File path: src/operator/numpy/random/np_exponential_op.h
##########
@@ -0,0 +1,147 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential_op.h
+ * \brief Operator for numpy sampling from exponential distribution.
+ */
+
+#ifndef MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+#define MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+
+#include <mxnet/operator_util.h>
+#include <algorithm>
+#include <string>
+#include <vector>
+#include <cmath>
+#include "../../elemwise_op_common.h"
+#include "../../mshadow_op.h"
+#include "../../mxnet_op.h"
+#include "../../operator_common.h"
+#include "../../tensor/elemwise_binary_broadcast_op.h"
+#include "./dist_common.h"
+
+namespace mxnet {
+namespace op {
+
+struct NumpyExponentialParam : public dmlc::Parameter<NumpyExponentialParam> {
+ dmlc::optional<float> scale;
+ dmlc::optional<mxnet::Tuple<int>> size;
+ DMLC_DECLARE_PARAMETER(NumpyExponentialParam) {
+ DMLC_DECLARE_FIELD(scale)
+ .set_default(dmlc::optional<float> (1.0));
+ DMLC_DECLARE_FIELD(size)
+ .set_default(dmlc::optional<mxnet::Tuple<int>>())
+ .describe("Output shape. If the given shape is, "
+ "e.g., (m, n, k), then m * n * k samples are drawn. "
+ "Default is None, in which case a single value is returned.");
+ }
+};
+
+template <typename DType>
+struct scalar_exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, float scale, float *threshold,
+ DType *out) {
+ out[i] = -scale * log(threshold[i]);
+ }
+};
+
+namespace mxnet_op {
+
+template <typename IType>
+struct check_legal_scale_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, IType *scalar, float* flag) {
+ if (scalar[i] < 0.0) {
+ flag[0] = -1.0;
+ }
+ }
+};
+
+
+template <int ndim, typename IType, typename OType>
+struct exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i,
+ const Shape<ndim> &stride,
+ const Shape<ndim> &oshape,
+ IType *scales, float* threshold, OType *out) {
+ Shape<ndim> coord = unravel(i, oshape);
+ auto idx = static_cast<index_t>(dot(coord, stride));
+ out[i] = -scales[idx] * log(threshold[i]);
+ }
+};
+
+} // namespace mxnet_op
+
+template <typename xpu>
+void NumpyExponentialForward(const nnvm::NodeAttrs &attrs,
+ const OpContext &ctx,
+ const std::vector<TBlob> &inputs,
+ const std::vector<OpReqType> &req,
+ const std::vector<TBlob> &outputs) {
+ using namespace mshadow;
+ using namespace mxnet_op;
+ const NumpyExponentialParam ¶m = nnvm::get<NumpyExponentialParam>(attrs.parsed);
+ Stream<xpu> *s = ctx.get_stream<xpu>();
+ index_t output_len = outputs[0].Size();
+ Random<xpu, float> *prnd = ctx.requested[0].get_random<xpu, float>(s);
+ Tensor<xpu, 1, float> workspace =
+ ctx.requested[1].get_space_typed<xpu, 1, float>(Shape1(output_len + 1), s);
+ Tensor<xpu, 1, float> uniform_tensor = workspace.Slice(0, output_len);
+ Tensor<xpu, 1, float> indicator_device = workspace.Slice(output_len, output_len + 1);
+ float indicator_host = 1.0;
+ float *indicator_device_ptr = indicator_device.dptr_;
+ Kernel<set_zero, xpu>::Launch(s, 1, indicator_device_ptr);
+ prnd->SampleUniform(&workspace, 0.0, 1.0);
+ if (param.scale.has_value()) {
+ CHECK_GE(param.scale.value(), 0.0) << "ValueError: expect scale >= 0";
+ MSHADOW_REAL_TYPE_SWITCH(outputs[0].type_flag_, DType, {
+ Kernel<scalar_exponential_kernel<DType>, xpu>::Launch(
+ s, outputs[0].Size(), param.scale.value(),
Review comment:
2-space indentation.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365679810
##########
File path: src/operator/numpy/random/np_exponential_op.cc
##########
@@ -0,0 +1,72 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential.cc
Review comment:
should be consistent with the actual file name.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365680048
##########
File path: src/operator/numpy/random/np_exponential_op.h
##########
@@ -0,0 +1,147 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential_op.h
+ * \brief Operator for numpy sampling from exponential distribution.
+ */
+
+#ifndef MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+#define MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+
+#include <mxnet/operator_util.h>
+#include <algorithm>
+#include <string>
+#include <vector>
+#include <cmath>
+#include "../../elemwise_op_common.h"
+#include "../../mshadow_op.h"
+#include "../../mxnet_op.h"
+#include "../../operator_common.h"
+#include "../../tensor/elemwise_binary_broadcast_op.h"
+#include "./dist_common.h"
+
+namespace mxnet {
+namespace op {
+
+struct NumpyExponentialParam : public dmlc::Parameter<NumpyExponentialParam> {
+ dmlc::optional<float> scale;
Review comment:
2-space indentation.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-mxnet] haojin2 commented on a change in pull request
#17280: [numpy] add op random.exponential
Posted by GitBox <gi...@apache.org>.
haojin2 commented on a change in pull request #17280: [numpy] add op random.exponential
URL: https://github.com/apache/incubator-mxnet/pull/17280#discussion_r365680112
##########
File path: src/operator/numpy/random/np_exponential_op.h
##########
@@ -0,0 +1,147 @@
+/*
+ * 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.
+ */
+
+/*!
+ * Copyright (c) 2019 by Contributors
+ * \file np_exponential_op.h
+ * \brief Operator for numpy sampling from exponential distribution.
+ */
+
+#ifndef MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+#define MXNET_OPERATOR_NUMPY_RANDOM_NP_EXPONENTIAL_OP_H_
+
+#include <mxnet/operator_util.h>
+#include <algorithm>
+#include <string>
+#include <vector>
+#include <cmath>
+#include "../../elemwise_op_common.h"
+#include "../../mshadow_op.h"
+#include "../../mxnet_op.h"
+#include "../../operator_common.h"
+#include "../../tensor/elemwise_binary_broadcast_op.h"
+#include "./dist_common.h"
+
+namespace mxnet {
+namespace op {
+
+struct NumpyExponentialParam : public dmlc::Parameter<NumpyExponentialParam> {
+ dmlc::optional<float> scale;
+ dmlc::optional<mxnet::Tuple<int>> size;
+ DMLC_DECLARE_PARAMETER(NumpyExponentialParam) {
+ DMLC_DECLARE_FIELD(scale)
+ .set_default(dmlc::optional<float> (1.0));
+ DMLC_DECLARE_FIELD(size)
+ .set_default(dmlc::optional<mxnet::Tuple<int>>())
+ .describe("Output shape. If the given shape is, "
+ "e.g., (m, n, k), then m * n * k samples are drawn. "
+ "Default is None, in which case a single value is returned.");
+ }
+};
+
+template <typename DType>
+struct scalar_exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, float scale, float *threshold,
+ DType *out) {
+ out[i] = -scale * log(threshold[i]);
+ }
+};
+
+namespace mxnet_op {
+
+template <typename IType>
+struct check_legal_scale_kernel {
+ MSHADOW_XINLINE static void Map(index_t i, IType *scalar, float* flag) {
+ if (scalar[i] < 0.0) {
+ flag[0] = -1.0;
+ }
+ }
+};
+
+
+template <int ndim, typename IType, typename OType>
+struct exponential_kernel {
+ MSHADOW_XINLINE static void Map(index_t i,
+ const Shape<ndim> &stride,
+ const Shape<ndim> &oshape,
+ IType *scales, float* threshold, OType *out) {
+ Shape<ndim> coord = unravel(i, oshape);
+ auto idx = static_cast<index_t>(dot(coord, stride));
+ out[i] = -scales[idx] * log(threshold[i]);
+ }
+};
+
+} // namespace mxnet_op
+
+template <typename xpu>
+void NumpyExponentialForward(const nnvm::NodeAttrs &attrs,
+ const OpContext &ctx,
Review comment:
alignment.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services