mindspore 2.2.14__cp39-cp39-win_amd64.whl → 2.4.0__cp39-cp39-win_amd64.whl

mindspore/train/summary/_summary_adapter.py

@@ -190,7 +190,7 @@ def _nptype_to_prototype(np_value):
     if proto is None:
         raise TypeError("Transform numpy type failed in Summary, expect numpy type is one of ['np.bool_', 'np.int8', "
                         "'np.int16', 'np.int32', 'np.int64', 'np.uint8', 'np.uint16', 'np.uint32', 'np.uint64', "
-                        "'np.float16', 'np.float', 'np.float64'].")
+                        "'np.float16', 'np.float_', 'np.float64'].")
 
     return proto
 

mindspore/train/summary/summary_record.py

@@ -21,13 +21,13 @@ import re
 import threading
 import time
 from collections import defaultdict
-import numpy as np
 
 from mindspore import log as logger
 from mindspore.nn import Cell
 from mindspore import context
 from mindspore._c_expression import security
 from mindspore._c_expression import Tensor as Tensor_
+from mindspore.common import dtype as mstype
 from mindspore.common.tensor import Tensor
 from mindspore import _checkparam as Validator
 from mindspore.common.api import _cell_graph_executor
@@ -37,7 +37,8 @@ from mindspore.train.summary._summary_adapter import get_event_file_name, packag
 from mindspore.train.summary._writer_pool import WriterPool
 from mindspore.train.summary.enums import PluginEnum
 from mindspore.ops.operations import debug_ops
-import mindspore.ops as ops
+from mindspore.ops.primitive import _run_op
+from mindspore.ops._primitive_cache import _get_cache_prim
 
 # for the moment, this lock is for caution's sake,
 # there are actually no any concurrences happening.
@@ -61,19 +62,12 @@ def _cache_summary_tensor_data(summary):
     with _summary_lock:
         for item in summary:
             SUMMARY_TENSOR_CACHE[item['name']] = item['data']
-    return True
+        return True
 
 
-def _get_summary_tensor_data(end_flag=None, del_end_flag=False):
+def _get_summary_tensor_data():
     """Get summary tensor data."""
     global SUMMARY_TENSOR_CACHE
-    if end_flag:
-        for _ in range(0, 100):
-            if SUMMARY_TENSOR_CACHE.get(end_flag):
-                break
-            time.sleep(0.01)
-        if del_end_flag and SUMMARY_TENSOR_CACHE.get(end_flag):
-            del SUMMARY_TENSOR_CACHE[end_flag]
     with _summary_lock:
         data = SUMMARY_TENSOR_CACHE
         SUMMARY_TENSOR_CACHE = {}
@@ -98,7 +92,7 @@ def _record_summary_tensor_data():
             "data": data[2]
         }
         summary_list.append(summary_value)
-    _ = _cache_summary_tensor_data(summary_list)
+    _cache_summary_tensor_data(summary_list)
     debug_ops.SUMMARY_TENSOR_CACHE = []
 
 
@@ -136,7 +130,7 @@ class SummaryRecord:
     SummaryRecord is used to record the summary data and lineage data.
 
     The API will create a summary file and lineage files lazily in a given directory and writes data to them.
-    It writes the data to files by executing the 'record' method. In addition to recording the data bubbled up from
+    It writes the data to files by executing the `record` method. In addition to recording the data bubbled up from
     the network by defining the summary operators, SummaryRecord also supports to record extra data which
     can be added by calling add_value.
 
@@ -217,7 +211,7 @@ class SummaryRecord:
         self._num_process = num_process
         self.raise_exception = raise_exception
         self._export_options = export_options
-        self.tensor_summary = ops.TensorSummary()
+        self._op_finished_init = False
 
         try:
             self._initialize()
@@ -287,25 +281,25 @@ class SummaryRecord:
                 LossLandscape]): The value to store.
 
                 - The data type of value should be 'GraphProto' (see `mindspore/ccsrc/anf_ir.proto
-                  <https://gitee.com/mindspore/mindspore/blob/r2.2/mindspore/ccsrc/utils/anf_ir.proto>`_) object
+                  <https://gitee.com/mindspore/mindspore/blob/master/mindspore/ccsrc/utils/anf_ir.proto>`_) object
                   when the plugin is 'graph'.
                 - The data type of value should be 'Tensor' object when the plugin is 'scalar', 'image', 'tensor'
                   or 'histogram'.
                 - The data type of value should be a 'TrainLineage' object when the plugin is 'train_lineage',
                   see `mindspore/ccsrc/lineage.proto
-                  <https://gitee.com/mindspore/mindspore/blob/r2.2/mindspore/ccsrc/utils/lineage.proto>`_.
+                  <https://gitee.com/mindspore/mindspore/blob/master/mindspore/ccsrc/utils/lineage.proto>`_.
                 - The data type of value should be a 'EvaluationLineage' object when the plugin is 'eval_lineage',
                   see `mindspore/ccsrc/lineage.proto
-                  <https://gitee.com/mindspore/mindspore/blob/r2.2/mindspore/ccsrc/utils/lineage.proto>`_.
+                  <https://gitee.com/mindspore/mindspore/blob/master/mindspore/ccsrc/utils/lineage.proto>`_.
                 - The data type of value should be a 'DatasetGraph' object when the plugin is 'dataset_graph',
                   see `mindspore/ccsrc/lineage.proto
-                  <https://gitee.com/mindspore/mindspore/blob/r2.2/mindspore/ccsrc/utils/lineage.proto>`_.
+                  <https://gitee.com/mindspore/mindspore/blob/master/mindspore/ccsrc/utils/lineage.proto>`_.
                 - The data type of value should be a 'UserDefinedInfo' object when the plugin is 'custom_lineage_data',
                   see `mindspore/ccsrc/lineage.proto
-                  <https://gitee.com/mindspore/mindspore/blob/r2.2/mindspore/ccsrc/utils/lineage.proto>`_.
+                  <https://gitee.com/mindspore/mindspore/blob/master/mindspore/ccsrc/utils/lineage.proto>`_.
                 - The data type of value should be a 'LossLandscape' object when the plugin is 'LANDSCAPE',
                   see `mindspore/ccsrc/summary.proto
-                  <https://gitee.com/mindspore/mindspore/blob/r2.2/mindspore/ccsrc/utils/summary.proto>`_.
+                  <https://gitee.com/mindspore/mindspore/blob/master/mindspore/ccsrc/utils/summary.proto>`_.
 
         Raises:
             ValueError: `plugin` is not in the optional value.
@@ -351,6 +345,39 @@ class SummaryRecord:
                              f', expect value is one of [tensor, scalar, image, histogram, train_lineage, '
                              f'eval_lineage, dataset_graph, custom_lineage_data, graph, landscape]')
 
+    def _sync_summary_data(self, step):
+        """
+        In the GE process, the summary receiving process is asynchronous.
+        When the record function is called, some data may not be received and
+        needs to be synchronized.
+        """
+        if context.get_context('device_target') != "Ascend":
+            return
+        if self._op_finished_init is False:
+            _ = Cell()
+            self._op_finished_init = True
+
+        step_end_flag = Tensor([1], dtype=mstype.float32)
+        image_step_end_flag = Tensor([[[[1]]]], dtype=mstype.float32)
+        _ = Cell()
+        name = "_step_end_flag_" + str(step)
+        tags = ("[:Tensor]", "[:Scalar]", "[:Image]", "[:Histogram]")
+        _run_op(_get_cache_prim(debug_ops.TensorSummary)(), "TensorSummary", (name, step_end_flag))
+        _run_op(_get_cache_prim(debug_ops.ScalarSummary)(), "ScalarSummary", (name, step_end_flag))
+        _run_op(_get_cache_prim(debug_ops.HistogramSummary)(), "HistogramSummary", (name, step_end_flag))
+        _run_op(_get_cache_prim(debug_ops.ImageSummary)(), "ImageSummary", (name, image_step_end_flag))
+        global SUMMARY_TENSOR_CACHE
+        for tag in tags:
+            item_name = name + tag
+            while item_name not in SUMMARY_TENSOR_CACHE:
+                time.sleep(0.004)
+
+        with _summary_lock:
+            for tag in tags:
+                item_name = name + tag
+                SUMMARY_TENSOR_CACHE.pop(item_name, None)
+
+
     def record(self, step, train_network=None, plugin_filter=None):
         r"""
         Record the summary.
@@ -367,7 +394,7 @@ class SummaryRecord:
 
         Raises:
             TypeError: `step` is not int, or `train_network` is not `mindspore.nn.Cell
-                <https://www.mindspore.cn/docs/en/r2.2/api_python/nn/mindspore.nn.Cell.html#mindspore-nn-cell>`_ .
+                <https://www.mindspore.cn/docs/en/master/api_python/nn/mindspore.nn.Cell.html#mindspore-nn-cell>`_ .
 
         Examples:
             >>> import mindspore as ms
@@ -399,8 +426,6 @@ class SummaryRecord:
                     return True
 
         if self._mode == 'train':
-            step_end_flag = Tensor((np.ones([1])).astype(np.int32))
-            self.tensor_summary("step_end_flag_" + str(step), step_end_flag)
             self._add_summary_tensor_data(step)
 
         if not plugin_filter:
@@ -453,16 +478,14 @@ class SummaryRecord:
                                         self._num_process,
                                         self.raise_exception,
                                         **filename_dict)
-        _ = _get_summary_tensor_data()
+        _get_summary_tensor_data()
         atexit.register(self.close)
 
     def _add_summary_tensor_data(self, step_index=-1):
         """Add summary tensor data."""
+        self._sync_summary_data(step_index)
         _record_summary_tensor_data()
-        end_flag = None
-        if step_index >= 0:
-            end_flag = "step_end_flag_" + str(step_index) + "[:Tensor]"
-        summary_data = _get_summary_tensor_data(end_flag=end_flag, del_end_flag=True)
+        summary_data = _get_summary_tensor_data()
         if not summary_data:
             logger.debug(f'No summary data bubbled from the network.')
         for name, tensor in summary_data.items():

mindspore/train/train_thor/convert_utils.py

@@ -175,7 +175,7 @@ class ConvertNetUtils:
             >>> from mindspore.nn import thor
             >>>
             >>> # Define the network structure of LeNet5. Refer to
-            >>> # https://gitee.com/mindspore/docs/blob/r2.2/docs/mindspore/code/lenet.py
+            >>> # https://gitee.com/mindspore/docs/blob/master/docs/mindspore/code/lenet.py
             >>> net = LeNet5()
             >>> temp = Tensor([4e-4, 1e-4, 1e-5, 1e-5], ms.float32)
             >>> opt = thor(net, learning_rate=temp, damping=temp, momentum=0.9, loss_scale=128, frequency=4)
@@ -239,10 +239,10 @@ class ConvertModelUtils:
             >>> from mindspore.nn import thor
             >>>
             >>> # Define the network structure of LeNet5. Refer to
-            >>> # https://gitee.com/mindspore/docs/blob/r2.2/docs/mindspore/code/lenet.py
+            >>> # https://gitee.com/mindspore/docs/blob/master/docs/mindspore/code/lenet.py
             >>> net = LeNet5()
             >>> # Create the dataset taking MNIST as an example. Refer to
-            >>> # https://gitee.com/mindspore/docs/blob/r2.2/docs/mindspore/code/mnist.py
+            >>> # https://gitee.com/mindspore/docs/blob/master/docs/mindspore/code/mnist.py
             >>> dataset = create_dataset()
             >>> temp = Tensor([4e-4, 1e-4, 1e-5, 1e-5], ms.float32)
             >>> opt = thor(net, learning_rate=temp, damping=temp, momentum=0.9, loss_scale=128, frequency=4)

mindspore/utils/__init__.py

@@ -0,0 +1,21 @@
+# Copyright 2024 Huawei Technologies Co., Ltd
+#
+# Licensed 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.
+# ============================================================================
+"""Utils module."""
+from __future__ import absolute_import
+from mindspore._c_expression import stress_detect
+from .utils import ExitByRequest
+
+# Symbols from utils module.
+__all__ = ["stress_detect", "ExitByRequest"]

mindspore/utils/utils.py

@@ -0,0 +1,60 @@
+# Copyright 2024 Huawei Technologies Co., Ltd
+#
+# Licensed 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.
+# ============================================================================
+"""mindspore utils."""
+from __future__ import absolute_import
+
+from mindspore.common import dtype as mstype
+from mindspore.common.tensor import Tensor
+from mindspore.ops import functional as F
+from mindspore.ops import operations as P
+from mindspore.context import ParallelMode
+from mindspore.parallel._utils import _get_parallel_mode
+from mindspore.common.api import jit_class
+
+
+@jit_class
+class ExitByRequest:
+    """
+    Gracefully exits the training process after get exit request.
+    """
+
+    def __init__(self):
+        super(ExitByRequest, self).__init__()
+        from mindspore.communication.management import get_group_size
+        self.all_reduce = P.AllReduce()
+        self.equal = P.Equal()
+        self.assign = P.Assign()
+        self.reduce_all = P.ReduceAll(keep_dims=False)
+        self.is_distributed = _get_parallel_mode() != ParallelMode.STAND_ALONE
+        if self.is_distributed:
+            self.group_size = get_group_size()
+            self.base = Tensor([self.group_size], dtype=mstype.int32)
+        self.base1 = Tensor([1], mstype.int32)
+        self.true = Tensor(True, mstype.bool_)
+
+    def exit_by_request(self, grad, init_value, exit_value):
+        """
+        update GracefulExit flag by Assign op, the value is the output of AllReduce op
+        :param grad: grad of net, or output of opt
+        :param init_value: input value of AllReduce, a parameter
+        :param exit_value: graceful exit value(out of AllReduce), update by Assign op
+        :return: grad
+        """
+        if self.is_distributed:
+            all_status = self.all_reduce(init_value)
+            equal = self.equal(all_status, self.base)
+            reduce_all = self.reduce_all(equal)
+            grad = F.depend(grad, self.assign(exit_value, reduce_all))
+        return grad

mindspore/version.py

@@ -1 +1 @@
-__version__ = '2.2.14'
+__version__ = '2.4.0'

{mindspore-2.2.14.dist-info → mindspore-2.4.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: mindspore
-Version: 2.2.14
+Version: 2.4.0
 Summary: MindSpore is a new open source deep learning training/inference framework that could be used for mobile, edge and cloud scenarios.
 Home-page: https://www.mindspore.cn
 Author: The MindSpore Authors
@@ -11,7 +11,7 @@ Project-URL: Sources, https://github.com/mindspore-ai/mindspore
 Project-URL: Issue Tracker, https://github.com/mindspore-ai/mindspore/issues
 Keywords: mindspore machine learning
 Platform: UNKNOWN
-Classifier: Development Status :: 4 - Beta
+Classifier: Development Status :: 5 - Production/Stable
 Classifier: Environment :: Console
 Classifier: Intended Audience :: Science/Research
 Classifier: Intended Audience :: Developers
@@ -19,6 +19,9 @@ Classifier: License :: OSI Approved :: Apache Software License
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Programming Language :: Python :: 3.7
 Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: C++
 Classifier: Topic :: Scientific/Engineering
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -27,7 +30,7 @@ Classifier: Topic :: Software Development :: Libraries
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.7
 Description-Content-Type: text/markdown
-Requires-Dist: numpy (>=1.17.0)
+Requires-Dist: numpy (<2.0.0,>=1.20.0)
 Requires-Dist: protobuf (>=3.13.0)
 Requires-Dist: asttokens (>=2.0.4)
 Requires-Dist: pillow (>=6.2.0)
@@ -35,6 +38,7 @@ Requires-Dist: scipy (>=1.5.4)
 Requires-Dist: packaging (>=20.0)
 Requires-Dist: psutil (>=5.6.1)
 Requires-Dist: astunparse (>=1.6.3)
+Requires-Dist: safetensors (>=0.4.0)
 
 ![MindSpore Logo](https://gitee.com/mindspore/mindspore/raw/master/docs/MindSpore-logo.png "MindSpore logo")
 
@@ -322,7 +326,7 @@ Project stable branches will be in one of the following states:
 | **r2.2**   | Maintained   | 2023-10-18               | Unmaintained <br> 2024-10-18 estimated |             |
 | **r2.1**   | Maintained   | 2023-07-29               | Unmaintained <br> 2024-07-29 estimated |             |
 | **r2.0**   | Maintained   | 2023-06-15               | Unmaintained <br> 2024-06-15 estimated |             |
-| **r1.10**  | Maintained   | 2023-02-02               | Unmaintained <br> 2024-02-02 estimated |             |
+| **r1.10**  | End Of Life  | 2023-02-02               |                                        | 2024-02-02  |
 | **r1.9**   | End Of Life  | 2022-10-26               |                                        | 2023-10-26  |
 | **r1.8**   | End Of Life  | 2022-07-29               |                                        | 2023-07-29  |
 | **r1.7**   | End Of Life  | 2022-04-29               |                                        | 2023-04-29  |