mindinsight.debugger.api.debugger_tensor 源代码

# Copyright 2021 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.
# ==============================================================================
"""DebuggerTensor."""
import re
from abc import ABC

import numpy as np

from mindinsight.debugger.common.log import LOGGER as log
from mindinsight.debugger.common.utils import NUMPY_TYPE_MAP
from mindinsight.debugger.stream_cache.data_loader import DumpTarget
from mindinsight.domain.graph.base import NodeType
from mindinsight.domain.graph.proto.ms_graph_pb2 import DataType


[文档]class DebuggerTensor(ABC): """ The tensor with specific rank, iteration and debugging info. .. warning:: All APIs in this class are experimental prototypes that are subject to change or deletion. Args: node (Node): The node that outputs this tensor. slot (int): The slot of the tensor on the node. iteration (int): The iteration of the tensor. Note: - Users should not instantiate this class manually. - The instances of this class is immutable. - A `DebuggerTensor` is always the output tensor of a node. """ def __init__(self, node, slot, iteration): self._node = node self._slot = slot self._iteration = iteration @property def node(self): """ Get the node that outputs this tensor. Returns: Node, the node that outputs this tensor. Examples: >>> from mindinsight.debugger import DumpAnalyzer >>> my_run = DumpAnalyzer(dump_dir="/path/to/your/dump_dir_with_dump_data") >>> tensors = list(my_run.select_tensors("conv")) >>> print(tensors[0].node) rank: 0 graph_name: kernel_graph_0 node_name: conv1.weight """ return self._node @property def slot(self): """ The output of the node may have several tensors. The slot refer to the index of the tensor Returns: int, the slot of the tensor on the node. Examples: >>> from mindinsight.debugger import DumpAnalyzer >>> my_run = DumpAnalyzer(dump_dir="/path/to/your/dump_dir_with_dump_data") >>> tensors = list(my_run.select_tensors("conv")) >>> print(tensors[0].slot) 0 """ return self._slot @property def iteration(self): """ Get iteration of the tensor. Returns: int, the iteration of the tensor. Examples: >>> from mindinsight.debugger import DumpAnalyzer >>> my_run = DumpAnalyzer(dump_dir="/path/to/your/dump_dir_with_dump_data") >>> tensors = list(my_run.select_tensors("conv")) >>> print(tensors[0].iteration) 0 """ return self._iteration @property def rank(self): """ The rank is the logical id of the device on which the tensor is generated. Returns: int, the rank for this tensor. Examples: >>> from mindinsight.debugger import DumpAnalyzer >>> my_run = DumpAnalyzer(dump_dir="/path/to/your/dump_dir_with_dump_data") >>> tensors = list(my_run.select_tensors("conv")) >>> print(tensors[0].rank) 0 """ return self._node.rank
[文档] def value(self): """ Get the value of the tensor. Returns: Union[numpy.array, None], The value could be None if failed to find data file in relative iteration. Examples: >>> from mindinsight.debugger import DumpAnalyzer >>> >>> def test_debugger_tensor(): ... my_run = DumpAnalyzer(dump_dir="/path/to/your/dump_dir_with_dump_data") ... tensors = list(my_run.select_tensors("conv")) ... # the tensors[0].value() maybe start the new process ... value = tensors[0].value() ... return value ... >>> if __name__ == "__main__": ... test_debugger_tensor() ... """ raise NotImplementedError
def __str__(self): feature = f"rank: {self.rank}\n" \ f"graph_name: {self.node.graph_name}\n" \ f"node_name: {self.node.name}\n" \ f"slot: {self.slot}\n" \ f"iteration: {self.iteration}" return feature
class DebuggerTensorImpl(DebuggerTensor): """DebuggerTensor implementation.""" @property def root_graph_id(self): """Get the root_graph_id for this tensor.""" return self._node.root_graph_id def has_value(self): """Check if the tensor has value.""" iteration = self.iteration if iteration is None: return False data_loader = self.node.debugger_engine.data_loader has_dump_output = bool(data_loader.dump_target in [DumpTarget.FULL, DumpTarget.OUTPUT_ONLY]) if not has_dump_output: return False if self.node.node_type == NodeType.CONSTANT: iteration = 'Constant' iter_dirs = data_loader.get_step_iter(rank_id=self.rank, step=iteration) file_found = self._file_found(iter_dirs) return file_found def _file_found(self, iter_dirs): """Check if the tensor file found in specified directory.""" node_name_without_scope = self.node.name.split('/')[-1] bin_pattern = node_name_without_scope + r".*.(\d+)$" npy_pattern = f"{node_name_without_scope}.*.output.{self.slot}.*.npy$" for iter_dir in iter_dirs: for tensor_path in iter_dir.iterdir(): file_name = tensor_path.name if re.search(bin_pattern, file_name) or re.search(npy_pattern, file_name): return True return False def value(self): if self.iteration is None: log.warning("The iteration of is not specified, no value returned.") return None base_node = self.node.base_node if hasattr(base_node, 'output') and hasattr(base_node.output, 'info'): info = base_node.output.info if isinstance(info, dict) and info.get("np_value") is not None: return info.get("np_value") debugger_engine = self.node.debugger_engine tensor_info = debugger_engine.dbg_services_module.TensorInfo( node_name=base_node.full_name if self.node.node_type == NodeType.CONSTANT else self.node.name, slot=self.slot, iteration=self.iteration, rank_id=self.rank, root_graph_id=self.root_graph_id, is_output=True) tensors = debugger_engine.dbg_service.read_tensors([tensor_info]) return self._to_numpy(tensors[0]) @staticmethod def _to_numpy(tensor_data): """Turn tensor data into Numpy.""" if tensor_data.data_size == 0: return None dtype_str = DataType.Name(tensor_data.dtype) np_type = NUMPY_TYPE_MAP.get(dtype_str) data = np.frombuffer(tensor_data.data_ptr, dtype=np_type) data = data.reshape(tensor_data.shape) return data