text
stringlengths
81
112k
Gets a new grouped symbol whose output contains inputs to output nodes of the original symbol. Example ------- >>> x = mx.sym.Variable('x') >>> y = mx.sym.Variable('y') >>> z = mx.sym.Variable('z') >>> a = y+z >>> b = x+a >>> b.get_children() <Symbol Grouped> >>> b.get_children().list_outputs() ['x', '_plus10_output'] >>> b.get_children().get_children().list_outputs() ['y', 'z'] Returns ------- sgroup : Symbol or None The children of the head node. If the symbol has no inputs then ``None`` will be returned. def get_children(self): """Gets a new grouped symbol whose output contains inputs to output nodes of the original symbol. Example ------- >>> x = mx.sym.Variable('x') >>> y = mx.sym.Variable('y') >>> z = mx.sym.Variable('z') >>> a = y+z >>> b = x+a >>> b.get_children() <Symbol Grouped> >>> b.get_children().list_outputs() ['x', '_plus10_output'] >>> b.get_children().get_children().list_outputs() ['y', 'z'] Returns ------- sgroup : Symbol or None The children of the head node. If the symbol has no inputs then ``None`` will be returned. """ handle = SymbolHandle() check_call(_LIB.MXSymbolGetChildren( self.handle, ctypes.byref(handle))) ret = Symbol(handle=handle) if len(ret.list_outputs()) == 0: return None return ret
Lists all the arguments in the symbol. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> c.list_arguments ['a', 'b'] Returns ------- args : list of string List containing the names of all the arguments required to compute the symbol. def list_arguments(self): """Lists all the arguments in the symbol. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> c.list_arguments ['a', 'b'] Returns ------- args : list of string List containing the names of all the arguments required to compute the symbol. """ size = ctypes.c_uint() sarr = ctypes.POINTER(ctypes.c_char_p)() check_call(_LIB.MXSymbolListArguments( self.handle, ctypes.byref(size), ctypes.byref(sarr))) return [py_str(sarr[i]) for i in range(size.value)]
Lists all the outputs in the symbol. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> c.list_outputs() ['_plus12_output'] Returns ------- list of str List of all the outputs. For most symbols, this list contains only the name of this symbol. For symbol groups, this is a list with the names of all symbols in the group. def list_outputs(self): """Lists all the outputs in the symbol. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> c.list_outputs() ['_plus12_output'] Returns ------- list of str List of all the outputs. For most symbols, this list contains only the name of this symbol. For symbol groups, this is a list with the names of all symbols in the group. """ size = ctypes.c_uint() sarr = ctypes.POINTER(ctypes.c_char_p)() check_call(_LIB.MXSymbolListOutputs( self.handle, ctypes.byref(size), ctypes.byref(sarr))) return [py_str(sarr[i]) for i in range(size.value)]
Lists all the auxiliary states in the symbol. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> c.list_auxiliary_states() [] Example of auxiliary states in `BatchNorm`. >>> data = mx.symbol.Variable('data') >>> weight = mx.sym.Variable(name='fc1_weight') >>> fc1 = mx.symbol.FullyConnected(data = data, weight=weight, name='fc1', num_hidden=128) >>> fc2 = mx.symbol.BatchNorm(fc1, name='batchnorm0') >>> fc2.list_auxiliary_states() ['batchnorm0_moving_mean', 'batchnorm0_moving_var'] Returns ------- aux_states : list of str List of the auxiliary states in input symbol. Notes ----- Auxiliary states are special states of symbols that do not correspond to an argument, and are not updated by gradient descent. Common examples of auxiliary states include the `moving_mean` and `moving_variance` in `BatchNorm`. Most operators do not have auxiliary states. def list_auxiliary_states(self): """Lists all the auxiliary states in the symbol. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> c.list_auxiliary_states() [] Example of auxiliary states in `BatchNorm`. >>> data = mx.symbol.Variable('data') >>> weight = mx.sym.Variable(name='fc1_weight') >>> fc1 = mx.symbol.FullyConnected(data = data, weight=weight, name='fc1', num_hidden=128) >>> fc2 = mx.symbol.BatchNorm(fc1, name='batchnorm0') >>> fc2.list_auxiliary_states() ['batchnorm0_moving_mean', 'batchnorm0_moving_var'] Returns ------- aux_states : list of str List of the auxiliary states in input symbol. Notes ----- Auxiliary states are special states of symbols that do not correspond to an argument, and are not updated by gradient descent. Common examples of auxiliary states include the `moving_mean` and `moving_variance` in `BatchNorm`. Most operators do not have auxiliary states. """ size = ctypes.c_uint() sarr = ctypes.POINTER(ctypes.c_char_p)() check_call(_LIB.MXSymbolListAuxiliaryStates( self.handle, ctypes.byref(size), ctypes.byref(sarr))) return [py_str(sarr[i]) for i in range(size.value)]
Lists all arguments and auxiliary states of this Symbol. Returns ------- inputs : list of str List of all inputs. Examples -------- >>> bn = mx.sym.BatchNorm(name='bn') >>> bn.list_arguments() ['bn_data', 'bn_gamma', 'bn_beta'] >>> bn.list_auxiliary_states() ['bn_moving_mean', 'bn_moving_var'] >>> bn.list_inputs() ['bn_data', 'bn_gamma', 'bn_beta', 'bn_moving_mean', 'bn_moving_var'] def list_inputs(self): """Lists all arguments and auxiliary states of this Symbol. Returns ------- inputs : list of str List of all inputs. Examples -------- >>> bn = mx.sym.BatchNorm(name='bn') >>> bn.list_arguments() ['bn_data', 'bn_gamma', 'bn_beta'] >>> bn.list_auxiliary_states() ['bn_moving_mean', 'bn_moving_var'] >>> bn.list_inputs() ['bn_data', 'bn_gamma', 'bn_beta', 'bn_moving_mean', 'bn_moving_var'] """ size = ctypes.c_uint() sarr = ctypes.POINTER(ctypes.c_char_p)() check_call(_LIB.NNSymbolListInputNames( self.handle, 0, ctypes.byref(size), ctypes.byref(sarr))) return [py_str(sarr[i]) for i in range(size.value)]
Infers the type of all arguments and all outputs, given the known types for some arguments. This function takes the known types of some arguments in either positional way or keyword argument way as input. It returns a tuple of `None` values if there is not enough information to deduce the missing types. Inconsistencies in the known types will cause an error to be raised. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> arg_types, out_types, aux_types = c.infer_type(a='float32') >>> arg_types [<type 'numpy.float32'>, <type 'numpy.float32'>] >>> out_types [<type 'numpy.float32'>] >>> aux_types [] Parameters ---------- *args : Type of known arguments in a positional way. Unknown type can be marked as None. **kwargs : Keyword arguments of known types. Returns ------- arg_types : list of numpy.dtype or None List of argument types. The order is same as the order of list_arguments(). out_types : list of numpy.dtype or None List of output types. The order is same as the order of list_outputs(). aux_types : list of numpy.dtype or None List of auxiliary state types. The order is same as the order of list_auxiliary_states(). def infer_type(self, *args, **kwargs): """Infers the type of all arguments and all outputs, given the known types for some arguments. This function takes the known types of some arguments in either positional way or keyword argument way as input. It returns a tuple of `None` values if there is not enough information to deduce the missing types. Inconsistencies in the known types will cause an error to be raised. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> arg_types, out_types, aux_types = c.infer_type(a='float32') >>> arg_types [<type 'numpy.float32'>, <type 'numpy.float32'>] >>> out_types [<type 'numpy.float32'>] >>> aux_types [] Parameters ---------- *args : Type of known arguments in a positional way. Unknown type can be marked as None. **kwargs : Keyword arguments of known types. Returns ------- arg_types : list of numpy.dtype or None List of argument types. The order is same as the order of list_arguments(). out_types : list of numpy.dtype or None List of output types. The order is same as the order of list_outputs(). aux_types : list of numpy.dtype or None List of auxiliary state types. The order is same as the order of list_auxiliary_states(). """ try: res = self._infer_type_impl(False, *args, **kwargs) if res[1] is None: arg_shapes, _, _ = self._infer_type_impl(True, *args, **kwargs) arg_names = self.list_arguments() unknowns = [] for name, dtype in zip(arg_names, arg_shapes): if not dtype: if len(unknowns) >= 10: unknowns.append('...') break unknowns.append('%s: %s' % (name, str(dtype))) warnings.warn( "Cannot decide type for the following arguments. " + "Consider providing them as input:\n\t" + "\n\t".join(unknowns), stacklevel=2) return res except MXNetError: print("infer_type error. Arguments:") for i, arg in enumerate(args): print(" #%d: %s" % (i, arg)) for k, v in kwargs.items(): print(" %s: %s" % (k, v)) raise
The actual implementation for calling type inference API. def _infer_type_impl(self, partial, *args, **kwargs): """The actual implementation for calling type inference API.""" # pylint: disable=too-many-locals if len(args) != 0 and len(kwargs) != 0: raise ValueError('Can only specify known argument \ types either by positional or kwargs way.') sdata = [] if len(args) != 0: keys = c_array(ctypes.c_char_p, []) for s in args: if s is not None: s = _numpy.dtype(s).type if s not in _DTYPE_NP_TO_MX: raise TypeError('Argument need to be one of ' + str(_DTYPE_NP_TO_MX)) sdata.append(_DTYPE_NP_TO_MX[s]) else: sdata.append(-1) else: str_keys = [] for k, v in kwargs.items(): v = _numpy.dtype(v).type if v in _DTYPE_NP_TO_MX: str_keys.append(k) sdata.append(_DTYPE_NP_TO_MX[v]) keys = c_str_array(str_keys) arg_type_size = mx_uint() arg_type_data = ctypes.POINTER(ctypes.c_int)() out_type_size = mx_uint() out_type_data = ctypes.POINTER(ctypes.c_int)() aux_type_size = mx_uint() aux_type_data = ctypes.POINTER(ctypes.c_int)() complete = ctypes.c_int() if partial: infer_func = _LIB.MXSymbolInferTypePartial else: infer_func = _LIB.MXSymbolInferType check_call(infer_func( self.handle, mx_uint(len(sdata)), keys, c_array_buf(ctypes.c_int, array('i', sdata)), ctypes.byref(arg_type_size), ctypes.byref(arg_type_data), ctypes.byref(out_type_size), ctypes.byref(out_type_data), ctypes.byref(aux_type_size), ctypes.byref(aux_type_data), ctypes.byref(complete))) if complete.value != 0: arg_types = [ _DTYPE_MX_TO_NP[arg_type_data[i]] for i in range(arg_type_size.value)] out_types = [ _DTYPE_MX_TO_NP[out_type_data[i]] for i in range(out_type_size.value)] aux_types = [ _DTYPE_MX_TO_NP[aux_type_data[i]] for i in range(aux_type_size.value)] return (arg_types, out_types, aux_types) else: return (None, None, None)
Infers the shapes of all arguments and all outputs given the known shapes of some arguments. This function takes the known shapes of some arguments in either positional way or keyword argument way as input. It returns a tuple of `None` values if there is not enough information to deduce the missing shapes. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> arg_shapes, out_shapes, aux_shapes = c.infer_shape(a=(3,3)) >>> arg_shapes [(3L, 3L), (3L, 3L)] >>> out_shapes [(3L, 3L)] >>> aux_shapes [] >>> c.infer_shape(a=(0,3)) # 0s in shape means unknown dimensions. So, returns None. (None, None, None) Inconsistencies in the known shapes will cause an error to be raised. See the following example: >>> data = mx.sym.Variable('data') >>> out = mx.sym.FullyConnected(data=data, name='fc1', num_hidden=1000) >>> out = mx.sym.Activation(data=out, act_type='relu') >>> out = mx.sym.FullyConnected(data=out, name='fc2', num_hidden=10) >>> weight_shape= (1, 100) >>> data_shape = (100, 100) >>> out.infer_shape(data=data_shape, fc1_weight=weight_shape) Error in operator fc1: Shape inconsistent, Provided=(1,100), inferred shape=(1000,100) Parameters ---------- *args : Shape of arguments in a positional way. Unknown shape can be marked as None. **kwargs : Keyword arguments of the known shapes. Returns ------- arg_shapes : list of tuple or None List of argument shapes. The order is same as the order of list_arguments(). out_shapes : list of tuple or None List of output shapes. The order is same as the order of list_outputs(). aux_shapes : list of tuple or None List of auxiliary state shapes. The order is same as the order of list_auxiliary_states(). def infer_shape(self, *args, **kwargs): """Infers the shapes of all arguments and all outputs given the known shapes of some arguments. This function takes the known shapes of some arguments in either positional way or keyword argument way as input. It returns a tuple of `None` values if there is not enough information to deduce the missing shapes. Example ------- >>> a = mx.sym.var('a') >>> b = mx.sym.var('b') >>> c = a + b >>> arg_shapes, out_shapes, aux_shapes = c.infer_shape(a=(3,3)) >>> arg_shapes [(3L, 3L), (3L, 3L)] >>> out_shapes [(3L, 3L)] >>> aux_shapes [] >>> c.infer_shape(a=(0,3)) # 0s in shape means unknown dimensions. So, returns None. (None, None, None) Inconsistencies in the known shapes will cause an error to be raised. See the following example: >>> data = mx.sym.Variable('data') >>> out = mx.sym.FullyConnected(data=data, name='fc1', num_hidden=1000) >>> out = mx.sym.Activation(data=out, act_type='relu') >>> out = mx.sym.FullyConnected(data=out, name='fc2', num_hidden=10) >>> weight_shape= (1, 100) >>> data_shape = (100, 100) >>> out.infer_shape(data=data_shape, fc1_weight=weight_shape) Error in operator fc1: Shape inconsistent, Provided=(1,100), inferred shape=(1000,100) Parameters ---------- *args : Shape of arguments in a positional way. Unknown shape can be marked as None. **kwargs : Keyword arguments of the known shapes. Returns ------- arg_shapes : list of tuple or None List of argument shapes. The order is same as the order of list_arguments(). out_shapes : list of tuple or None List of output shapes. The order is same as the order of list_outputs(). aux_shapes : list of tuple or None List of auxiliary state shapes. The order is same as the order of list_auxiliary_states(). """ try: res = self._infer_shape_impl(False, *args, **kwargs) if res[1] is None: arg_shapes, _, _ = self._infer_shape_impl(True, *args, **kwargs) arg_names = self.list_arguments() unknowns = [] for name, shape in zip(arg_names, arg_shapes): if is_np_compat(): shape_is_none = not shape or -1 in shape else: shape_is_none = not shape or 0 in shape if shape_is_none: if len(unknowns) >= 10: unknowns.append('...') break unknowns.append('%s: %s' % (name, str(shape))) warnings.warn( "Cannot decide shape for the following arguments " + "(0s in shape means unknown dimensions). " + "Consider providing them as input:\n\t" + "\n\t".join(unknowns), stacklevel=2) return res except MXNetError: print("infer_shape error. Arguments:") for i, arg in enumerate(args): print(" #%d: %s" % (i, arg)) for k, v in kwargs.items(): print(" %s: %s" % (k, v)) raise
The actual implementation for calling shape inference API. def _infer_shape_impl(self, partial, *args, **kwargs): """The actual implementation for calling shape inference API.""" # pylint: disable=too-many-locals if len(args) != 0 and len(kwargs) != 0: raise ValueError('Can only specify known argument \ shapes either by positional or kwargs way.') sdata = [] indptr = [0] if len(args) != 0: keys = c_array(ctypes.c_char_p, []) for i, s in enumerate(args): if s is not None: if not isinstance(s, tuple): raise TypeError("Arguments need to be shapes (tuple), " "but argument %d is %s." % (i, type(s))) sdata.extend(s) indptr.append(len(sdata)) else: str_keys = [] for k, v in kwargs.items(): if not isinstance(v, tuple): raise TypeError("Arguments need to be shapes (tuple), " "but '%s' is %s." % (k, type(v))) str_keys.append(k) sdata.extend(v) indptr.append(len(sdata)) keys = c_str_array(str_keys) arg_shape_size = mx_uint() arg_shape_ndim = ctypes.POINTER(mx_int)() arg_shape_data = ctypes.POINTER(ctypes.POINTER(mx_int))() out_shape_size = mx_uint() out_shape_ndim = ctypes.POINTER(mx_int)() out_shape_data = ctypes.POINTER(ctypes.POINTER(mx_int))() aux_shape_size = mx_uint() aux_shape_ndim = ctypes.POINTER(mx_int)() aux_shape_data = ctypes.POINTER(ctypes.POINTER(mx_int))() complete = ctypes.c_int() if partial: infer_func = _LIB.MXSymbolInferShapePartialEx else: infer_func = _LIB.MXSymbolInferShapeEx check_call(infer_func( self.handle, mx_uint(len(indptr) - 1), keys, c_array_buf(mx_uint, array('I', indptr)), c_array_buf(mx_int, array('i', sdata)), ctypes.byref(arg_shape_size), ctypes.byref(arg_shape_ndim), ctypes.byref(arg_shape_data), ctypes.byref(out_shape_size), ctypes.byref(out_shape_ndim), ctypes.byref(out_shape_data), ctypes.byref(aux_shape_size), ctypes.byref(aux_shape_ndim), ctypes.byref(aux_shape_data), ctypes.byref(complete))) if complete.value != 0: arg_shapes = [tuple(arg_shape_data[i][:arg_shape_ndim[i]]) if arg_shape_ndim[i] >= 0 else None for i in range(arg_shape_size.value)] out_shapes = [tuple(out_shape_data[i][:out_shape_ndim[i]]) if out_shape_ndim[i] >= 0 else None for i in range(out_shape_size.value)] aux_shapes = [tuple(aux_shape_data[i][:aux_shape_ndim[i]]) if aux_shape_ndim[i] >= 0 else None for i in range(aux_shape_size.value)] return (arg_shapes, out_shapes, aux_shapes) else: return (None, None, None)
Saves symbol to a file. You can also use pickle to do the job if you only work on python. The advantage of `load`/`save` functions is that the file contents are language agnostic. This means the model saved by one language binding can be loaded by a different language binding of `MXNet`. You also get the benefit of being able to directly load/save from cloud storage(S3, HDFS). Parameters ---------- fname : str The name of the file. - "s3://my-bucket/path/my-s3-symbol" - "hdfs://my-bucket/path/my-hdfs-symbol" - "/path-to/my-local-symbol" See Also -------- symbol.load : Used to load symbol from file. def save(self, fname): """Saves symbol to a file. You can also use pickle to do the job if you only work on python. The advantage of `load`/`save` functions is that the file contents are language agnostic. This means the model saved by one language binding can be loaded by a different language binding of `MXNet`. You also get the benefit of being able to directly load/save from cloud storage(S3, HDFS). Parameters ---------- fname : str The name of the file. - "s3://my-bucket/path/my-s3-symbol" - "hdfs://my-bucket/path/my-hdfs-symbol" - "/path-to/my-local-symbol" See Also -------- symbol.load : Used to load symbol from file. """ if not isinstance(fname, string_types): raise TypeError('fname need to be string') check_call(_LIB.MXSymbolSaveToFile(self.handle, c_str(fname)))
Saves symbol to a JSON string. See Also -------- symbol.load_json : Used to load symbol from JSON string. def tojson(self): """Saves symbol to a JSON string. See Also -------- symbol.load_json : Used to load symbol from JSON string. """ json_str = ctypes.c_char_p() check_call(_LIB.MXSymbolSaveToJSON(self.handle, ctypes.byref(json_str))) return py_str(json_str.value)
Helper function to get NDArray lists handles from various inputs. Parameters ---------- arg_key : str The name of argument, used for error message. args : list of NDArray or dict of str to NDArray Input arguments to the symbols. If type is list of NDArray, the position is in the same order of arg_names. If type is dict of str to NDArray, then it maps the name of arguments to the corresponding NDArray, args_names : list of string List of argument names. allow_missing : boolean Whether missing argument is allowed. When allowed, the missing handle will be set to None(null) Returns ------- handles : list of NDArrayHandle The positional list of NDArrayHandles generated from input. def _get_ndarray_inputs(arg_key, args, arg_names, allow_missing): """Helper function to get NDArray lists handles from various inputs. Parameters ---------- arg_key : str The name of argument, used for error message. args : list of NDArray or dict of str to NDArray Input arguments to the symbols. If type is list of NDArray, the position is in the same order of arg_names. If type is dict of str to NDArray, then it maps the name of arguments to the corresponding NDArray, args_names : list of string List of argument names. allow_missing : boolean Whether missing argument is allowed. When allowed, the missing handle will be set to None(null) Returns ------- handles : list of NDArrayHandle The positional list of NDArrayHandles generated from input. """ # setup args arg_handles = [] arg_arrays = [] if isinstance(args, list): if len(args) != len(arg_names): raise ValueError('Length of %s does not match the number of arguments' % arg_key) for narr in args: if narr is None and allow_missing: arg_handles.append(None) elif not isinstance(narr, NDArray): raise TypeError('Only accept list of NDArrays or dict of str to NDArray') else: arg_handles.append(narr.handle) arg_arrays = args elif isinstance(args, dict): for name in arg_names: if name in args: narr = args[name] if not isinstance(narr, NDArray): raise TypeError('Only accept list of NDArrays or dict of str to NDArray') arg_handles.append(narr.handle) arg_arrays.append(narr) else: if allow_missing: arg_handles.append(None) arg_arrays.append(None) else: raise ValueError('key `%s` is missing in `%s`' % (name, arg_key)) else: raise TypeError('Only accept list of NDArrays or dict of str to NDArray') return c_array(NDArrayHandle, arg_handles), arg_arrays
Bind current symbol to get an executor, allocate all the arguments needed. Allows specifying data types. This function simplifies the binding procedure. You need to specify only input data shapes. Before binding the executor, the function allocates arguments and auxiliary states that were not explicitly specified. Allows specifying data types. Example ------- >>> x = mx.sym.Variable('x') >>> y = mx.sym.FullyConnected(x, num_hidden=4) >>> exe = y.simple_bind(mx.cpu(), x=(5,4), grad_req='null') >>> exe.forward() [<NDArray 5x4 @cpu(0)>] >>> exe.outputs[0].asnumpy() array([[ 0., 0., 0., 0.], [ 0., 0., 0., 0.], [ 0., 0., 0., 0.], [ 0., 0., 0., 0.], [ 0., 0., 0., 0.]], dtype=float32) >>> exe.arg_arrays [<NDArray 5x4 @cpu(0)>, <NDArray 4x4 @cpu(0)>, <NDArray 4 @cpu(0)>] >>> exe.grad_arrays [<NDArray 5x4 @cpu(0)>, <NDArray 4x4 @cpu(0)>, <NDArray 4 @cpu(0)>] Parameters ---------- ctx : Context The device context the generated executor to run on. grad_req: string {'write', 'add', 'null'}, or list of str or dict of str to str, optional To specify how we should update the gradient to the `args_grad`. - 'write' means every time gradient is written to specified `args_grad` NDArray. - 'add' means every time gradient is added to the specified NDArray. - 'null' means no action is taken, the gradient may not be calculated. type_dict : Dict of str->numpy.dtype Input type dictionary, name->dtype stype_dict : Dict of str->str Input storage type dictionary, name->storage_type group2ctx : Dict of string to mx.Context The dict mapping the `ctx_group` attribute to the context assignment. shared_arg_names : List of string The argument names whose `NDArray` of shared_exec can be reused for initializing the current executor. shared_exec : Executor The executor whose arg_arrays, arg_arrays, grad_arrays, and aux_arrays can be reused for initializing the current executor. shared_buffer : Dict of string to `NDArray` The dict mapping argument names to the `NDArray` that can be reused for initializing the current executor. This buffer will be checked for reuse if one argument name of the current executor is not found in `shared_arg_names`. The `NDArray` s are expected have default storage type. kwargs : Dict of str->shape Input shape dictionary, name->shape Returns ------- executor : mxnet.Executor The generated executor def simple_bind(self, ctx, grad_req='write', type_dict=None, stype_dict=None, group2ctx=None, shared_arg_names=None, shared_exec=None, shared_buffer=None, **kwargs): """Bind current symbol to get an executor, allocate all the arguments needed. Allows specifying data types. This function simplifies the binding procedure. You need to specify only input data shapes. Before binding the executor, the function allocates arguments and auxiliary states that were not explicitly specified. Allows specifying data types. Example ------- >>> x = mx.sym.Variable('x') >>> y = mx.sym.FullyConnected(x, num_hidden=4) >>> exe = y.simple_bind(mx.cpu(), x=(5,4), grad_req='null') >>> exe.forward() [<NDArray 5x4 @cpu(0)>] >>> exe.outputs[0].asnumpy() array([[ 0., 0., 0., 0.], [ 0., 0., 0., 0.], [ 0., 0., 0., 0.], [ 0., 0., 0., 0.], [ 0., 0., 0., 0.]], dtype=float32) >>> exe.arg_arrays [<NDArray 5x4 @cpu(0)>, <NDArray 4x4 @cpu(0)>, <NDArray 4 @cpu(0)>] >>> exe.grad_arrays [<NDArray 5x4 @cpu(0)>, <NDArray 4x4 @cpu(0)>, <NDArray 4 @cpu(0)>] Parameters ---------- ctx : Context The device context the generated executor to run on. grad_req: string {'write', 'add', 'null'}, or list of str or dict of str to str, optional To specify how we should update the gradient to the `args_grad`. - 'write' means every time gradient is written to specified `args_grad` NDArray. - 'add' means every time gradient is added to the specified NDArray. - 'null' means no action is taken, the gradient may not be calculated. type_dict : Dict of str->numpy.dtype Input type dictionary, name->dtype stype_dict : Dict of str->str Input storage type dictionary, name->storage_type group2ctx : Dict of string to mx.Context The dict mapping the `ctx_group` attribute to the context assignment. shared_arg_names : List of string The argument names whose `NDArray` of shared_exec can be reused for initializing the current executor. shared_exec : Executor The executor whose arg_arrays, arg_arrays, grad_arrays, and aux_arrays can be reused for initializing the current executor. shared_buffer : Dict of string to `NDArray` The dict mapping argument names to the `NDArray` that can be reused for initializing the current executor. This buffer will be checked for reuse if one argument name of the current executor is not found in `shared_arg_names`. The `NDArray` s are expected have default storage type. kwargs : Dict of str->shape Input shape dictionary, name->shape Returns ------- executor : mxnet.Executor The generated executor """ # data types num_provided_arg_types = 0 provided_arg_type_names = ctypes.POINTER(ctypes.c_char_p)() # provided type argument names provided_arg_type_data = ctypes.POINTER(mx_uint)() # provided types if type_dict is not None: provided_arg_type_names = [] provided_arg_type_data = [] for k, v in type_dict.items(): v = _numpy.dtype(v).type if v in _DTYPE_NP_TO_MX: provided_arg_type_names.append(k) provided_arg_type_data.append(_DTYPE_NP_TO_MX[v]) num_provided_arg_types = mx_uint(len(provided_arg_type_names)) provided_arg_type_names = c_str_array(provided_arg_type_names) provided_arg_type_data = c_array_buf(ctypes.c_int, array('i', provided_arg_type_data)) # storage types num_provided_arg_stypes = 0 # provided storage type argument names provided_arg_stype_names = ctypes.POINTER(ctypes.c_char_p)() provided_arg_stype_data = ctypes.POINTER(mx_uint)() # provided storage types if stype_dict is not None: provided_arg_stype_names = [] provided_arg_stype_data = [] for k, v in stype_dict.items(): if v in _STORAGE_TYPE_STR_TO_ID: provided_arg_stype_names.append(k) provided_arg_stype_data.append(_STORAGE_TYPE_STR_TO_ID[v]) num_provided_arg_stypes = mx_uint(len(provided_arg_stype_names)) provided_arg_stype_names = c_str_array(provided_arg_stype_names) provided_arg_stype_data = c_array_buf(ctypes.c_int, array('i', provided_arg_stype_data)) provided_arg_shape_data = [] # shape data # argument shape index in sdata, # e.g. [sdata[indptr[0]], sdata[indptr[1]]) is the shape of the first arg provided_arg_shape_idx = [0] provided_arg_shape_names = [] # provided argument names for k, v in kwargs.items(): # if k not in listed_arguments and k not in listed_aux_states: # raise ValueError('arg name %s is not valid', k) if isinstance(v, tuple): provided_arg_shape_names.append(k) provided_arg_shape_data.extend(v) provided_arg_shape_idx.append(len(provided_arg_shape_data)) provided_req_type_list_len = 0 provided_grad_req_types = ctypes.POINTER(ctypes.c_char_p)() provided_grad_req_names = ctypes.POINTER(ctypes.c_char_p)() if grad_req is not None: if isinstance(grad_req, string_types): # use provided_req_type_list_len = 0 to indicate this situation provided_req_type_list_len = 0 provided_grad_req_types = [grad_req] elif isinstance(grad_req, list): if len(grad_req) == 0: raise RuntimeError('grad_req in simple_bind cannot be an empty list') provided_grad_req_types = grad_req provided_req_type_list_len = len(provided_grad_req_types) elif isinstance(grad_req, dict): if len(grad_req) == 0: raise RuntimeError('grad_req in simple_bind cannot be an empty dict') provided_grad_req_names = [] provided_grad_req_types = [] for k, v in grad_req.items(): provided_grad_req_names.append(k) provided_grad_req_types.append(v) provided_grad_req_names = c_str_array(provided_grad_req_names) provided_req_type_list_len = len(provided_grad_req_types) provided_grad_req_types = c_str_array(provided_grad_req_types) num_ctx_map_keys = mx_uint(0) ctx_map_keys = ctypes.POINTER(ctypes.c_char_p)() ctx_map_dev_types = ctypes.POINTER(ctypes.c_int)() ctx_map_dev_ids = ctypes.POINTER(ctypes.c_int)() if group2ctx is not None: ctx_map_keys = [] ctx_map_dev_types = [] ctx_map_dev_ids = [] for key, val in group2ctx.items(): ctx_map_keys.append(key) ctx_map_dev_types.append(val.device_typeid) ctx_map_dev_ids.append(val.device_id) num_ctx_map_keys = mx_uint(len(ctx_map_keys)) ctx_map_keys = c_str_array(ctx_map_keys) ctx_map_dev_types = c_array(ctypes.c_int, array('i', ctx_map_dev_types)) ctx_map_dev_ids = c_array(ctypes.c_int, array('i', ctx_map_dev_ids)) # prepare param names shared_arg_name_list = [] if shared_arg_names is not None: if not isinstance(shared_arg_names, list): raise ValueError('shared_arg_names in simple_bind must be a list or None') shared_arg_name_list = shared_arg_names # prepare shared_buffer if shared_buffer is None: shared_buffer_len = ctypes.c_int(-1) shared_buffer_names = ctypes.POINTER(ctypes.c_char_p)() shared_buffer_handles = ctypes.POINTER(NDArrayHandle)() else: if not isinstance(shared_buffer, dict): raise ValueError('shared_buffer in simple_bind must be dict or None') buffer_names = shared_buffer.keys() buffer_arrays = shared_buffer.values() for v in buffer_arrays: assert(v.stype == 'default'), \ "shared_buffer is expected to only contain NDArrays with default storage" shared_buffer_names = c_str_array(buffer_names) shared_buffer_len = ctypes.c_int(len(buffer_arrays)) shared_buffer_handles = c_handle_array(buffer_arrays) updated_shared_buffer_names = ctypes.POINTER(ctypes.c_char_p)() updated_shared_buffer_handles = ctypes.POINTER(NDArrayHandle)() # prepare shared_exec_handle shared_exec_handle = shared_exec.handle if shared_exec is not None else ExecutorHandle() # prepare current executor handle exe_handle = ExecutorHandle() # prepare current executor's in_args, arg_grads, and aux_states num_in_args = ctypes.c_uint() in_arg_handles = ctypes.POINTER(NDArrayHandle)() arg_grad_handles = ctypes.POINTER(NDArrayHandle)() num_aux_states = ctypes.c_uint() aux_state_handles = ctypes.POINTER(NDArrayHandle)() try: check_call(_LIB.MXExecutorSimpleBindEx(self.handle, ctypes.c_int(ctx.device_typeid), ctypes.c_int(ctx.device_id), num_ctx_map_keys, ctx_map_keys, ctx_map_dev_types, ctx_map_dev_ids, mx_uint(provided_req_type_list_len), provided_grad_req_names, provided_grad_req_types, mx_uint(len(provided_arg_shape_names)), c_str_array(provided_arg_shape_names), c_array_buf(mx_int, array('I', provided_arg_shape_data)), c_array_buf(mx_uint, array('i', provided_arg_shape_idx)), num_provided_arg_types, provided_arg_type_names, provided_arg_type_data, num_provided_arg_stypes, provided_arg_stype_names, provided_arg_stype_data, mx_uint(len(shared_arg_name_list)), c_str_array(shared_arg_name_list), ctypes.byref(shared_buffer_len), shared_buffer_names, shared_buffer_handles, ctypes.byref(updated_shared_buffer_names), ctypes.byref(updated_shared_buffer_handles), ctypes.byref(num_in_args), ctypes.byref(in_arg_handles), ctypes.byref(arg_grad_handles), ctypes.byref(num_aux_states), ctypes.byref(aux_state_handles), shared_exec_handle, ctypes.byref(exe_handle))) except MXNetError as e: error_msg = "simple_bind error. Arguments:\n" for k, v in kwargs.items(): error_msg += "%s: %s\n" % (k, v) error_msg += "%s" % e raise RuntimeError(error_msg) # update shared_buffer if shared_buffer is not None: for i in range(shared_buffer_len.value): k = py_str(updated_shared_buffer_names[i]) v = NDArray(NDArrayHandle(updated_shared_buffer_handles[i])) shared_buffer[k] = v # create in_args, arg_grads, and aux_states for the current executor arg_arrays = [_ndarray_cls(NDArrayHandle(in_arg_handles[i])) for i in range(num_in_args.value)] grad_arrays = [_ndarray_cls(NDArrayHandle(arg_grad_handles[i])) if arg_grad_handles[i] is not None else None for i in range(num_in_args.value)] aux_arrays = [_ndarray_cls(NDArrayHandle(aux_state_handles[i])) for i in range(num_aux_states.value)] executor = Executor(exe_handle, self, ctx, grad_req, group2ctx) executor.arg_arrays = arg_arrays executor.grad_arrays = grad_arrays executor.aux_arrays = aux_arrays return executor
Binds the current symbol to an executor and returns it. We first declare the computation and then bind to the data to run. This function returns an executor which provides method `forward()` method for evaluation and a `outputs()` method to get all the results. Example ------- >>> a = mx.sym.Variable('a') >>> b = mx.sym.Variable('b') >>> c = a + b <Symbol _plus1> >>> ex = c.bind(ctx=mx.cpu(), args={'a' : mx.nd.ones([2,3]), 'b' : mx.nd.ones([2,3])}) >>> ex.forward() [<NDArray 2x3 @cpu(0)>] >>> ex.outputs[0].asnumpy() [[ 2. 2. 2.] [ 2. 2. 2.]] Parameters ---------- ctx : Context The device context the generated executor to run on. args : list of NDArray or dict of str to NDArray Input arguments to the symbol. - If the input type is a list of `NDArray`, the order should be same as the order of `list_arguments()`. - If the input type is a dict of str to `NDArray`, then it maps the name of arguments to the corresponding `NDArray`. - In either case, all the arguments must be provided. args_grad : list of NDArray or dict of str to `NDArray`, optional When specified, `args_grad` provides NDArrays to hold the result of gradient value in backward. - If the input type is a list of `NDArray`, the order should be same as the order of `list_arguments()`. - If the input type is a dict of str to `NDArray`, then it maps the name of arguments to the corresponding NDArray. - When the type is a dict of str to `NDArray`, one only need to provide the dict for required argument gradient. Only the specified argument gradient will be calculated. grad_req : {'write', 'add', 'null'}, or list of str or dict of str to str, optional To specify how we should update the gradient to the `args_grad`. - 'write' means everytime gradient is write to specified `args_grad` `NDArray`. - 'add' means everytime gradient is add to the specified NDArray. - 'null' means no action is taken, the gradient may not be calculated. aux_states : list of `NDArray`, or dict of str to `NDArray`, optional Input auxiliary states to the symbol, only needed when the output of `list_auxiliary_states()` is not empty. - If the input type is a list of `NDArray`, the order should be same as the order of `list_auxiliary_states()`. - If the input type is a dict of str to `NDArray`, then it maps the name of `auxiliary_states` to the corresponding `NDArray`, - In either case, all the auxiliary states need to be provided. group2ctx : Dict of string to mx.Context The dict mapping the `ctx_group` attribute to the context assignment. shared_exec : mx.executor.Executor Executor to share memory with. This is intended for runtime reshaping, variable length sequences, etc. The returned executor shares state with `shared_exec`, and should not be used in parallel with it. Returns ------- executor : Executor The generated executor Notes ----- Auxiliary states are the special states of symbols that do not correspond to an argument, and do not have gradient but are still useful for the specific operations. Common examples of auxiliary states include the `moving_mean` and `moving_variance` states in `BatchNorm`. Most operators do not have auxiliary states and in those cases, this parameter can be safely ignored. One can give up gradient by using a dict in `args_grad` and only specify gradient they interested in. def bind(self, ctx, args, args_grad=None, grad_req='write', aux_states=None, group2ctx=None, shared_exec=None): """Binds the current symbol to an executor and returns it. We first declare the computation and then bind to the data to run. This function returns an executor which provides method `forward()` method for evaluation and a `outputs()` method to get all the results. Example ------- >>> a = mx.sym.Variable('a') >>> b = mx.sym.Variable('b') >>> c = a + b <Symbol _plus1> >>> ex = c.bind(ctx=mx.cpu(), args={'a' : mx.nd.ones([2,3]), 'b' : mx.nd.ones([2,3])}) >>> ex.forward() [<NDArray 2x3 @cpu(0)>] >>> ex.outputs[0].asnumpy() [[ 2. 2. 2.] [ 2. 2. 2.]] Parameters ---------- ctx : Context The device context the generated executor to run on. args : list of NDArray or dict of str to NDArray Input arguments to the symbol. - If the input type is a list of `NDArray`, the order should be same as the order of `list_arguments()`. - If the input type is a dict of str to `NDArray`, then it maps the name of arguments to the corresponding `NDArray`. - In either case, all the arguments must be provided. args_grad : list of NDArray or dict of str to `NDArray`, optional When specified, `args_grad` provides NDArrays to hold the result of gradient value in backward. - If the input type is a list of `NDArray`, the order should be same as the order of `list_arguments()`. - If the input type is a dict of str to `NDArray`, then it maps the name of arguments to the corresponding NDArray. - When the type is a dict of str to `NDArray`, one only need to provide the dict for required argument gradient. Only the specified argument gradient will be calculated. grad_req : {'write', 'add', 'null'}, or list of str or dict of str to str, optional To specify how we should update the gradient to the `args_grad`. - 'write' means everytime gradient is write to specified `args_grad` `NDArray`. - 'add' means everytime gradient is add to the specified NDArray. - 'null' means no action is taken, the gradient may not be calculated. aux_states : list of `NDArray`, or dict of str to `NDArray`, optional Input auxiliary states to the symbol, only needed when the output of `list_auxiliary_states()` is not empty. - If the input type is a list of `NDArray`, the order should be same as the order of `list_auxiliary_states()`. - If the input type is a dict of str to `NDArray`, then it maps the name of `auxiliary_states` to the corresponding `NDArray`, - In either case, all the auxiliary states need to be provided. group2ctx : Dict of string to mx.Context The dict mapping the `ctx_group` attribute to the context assignment. shared_exec : mx.executor.Executor Executor to share memory with. This is intended for runtime reshaping, variable length sequences, etc. The returned executor shares state with `shared_exec`, and should not be used in parallel with it. Returns ------- executor : Executor The generated executor Notes ----- Auxiliary states are the special states of symbols that do not correspond to an argument, and do not have gradient but are still useful for the specific operations. Common examples of auxiliary states include the `moving_mean` and `moving_variance` states in `BatchNorm`. Most operators do not have auxiliary states and in those cases, this parameter can be safely ignored. One can give up gradient by using a dict in `args_grad` and only specify gradient they interested in. """ # pylint: disable=too-many-locals, too-many-branches if not isinstance(ctx, Context): raise TypeError("Context type error") listed_arguments = self.list_arguments() args_handle, args = self._get_ndarray_inputs('args', args, listed_arguments, False) # setup args gradient if args_grad is None: args_grad_handle = c_array(NDArrayHandle, [None] * len(args)) else: args_grad_handle, args_grad = self._get_ndarray_inputs( 'args_grad', args_grad, listed_arguments, True) if aux_states is None: aux_states = [] aux_args_handle, aux_states = self._get_ndarray_inputs( 'aux_states', aux_states, self.list_auxiliary_states(), False) # setup requirements if isinstance(grad_req, string_types): if grad_req not in _GRAD_REQ_MAP: raise ValueError('grad_req must be in %s' % str(_GRAD_REQ_MAP)) reqs_array = c_array_buf(mx_uint, array('I', [_GRAD_REQ_MAP[grad_req]] * len(listed_arguments))) elif isinstance(grad_req, list): reqs_array = c_array_buf(mx_uint, array('I', [_GRAD_REQ_MAP[item] for item in grad_req])) elif isinstance(grad_req, dict): req_array = [] for name in listed_arguments: if name in grad_req: req_array.append(_GRAD_REQ_MAP[grad_req[name]]) else: req_array.append(0) reqs_array = c_array_buf(mx_uint, array('I', req_array)) ctx_map_keys = [] ctx_map_dev_types = [] ctx_map_dev_ids = [] if group2ctx: for key, val in group2ctx.items(): ctx_map_keys.append(key) ctx_map_dev_types.append(val.device_typeid) ctx_map_dev_ids.append(val.device_id) handle = ExecutorHandle() shared_handle = shared_exec.handle if shared_exec is not None else ExecutorHandle() check_call(_LIB.MXExecutorBindEX(self.handle, ctypes.c_int(ctx.device_typeid), ctypes.c_int(ctx.device_id), mx_uint(len(ctx_map_keys)), c_str_array(ctx_map_keys), c_array_buf(ctypes.c_int, array('i', ctx_map_dev_types)), c_array_buf(ctypes.c_int, array('i', ctx_map_dev_ids)), mx_uint(len(args)), args_handle, args_grad_handle, reqs_array, mx_uint(len(aux_states)), aux_args_handle, shared_handle, ctypes.byref(handle))) executor = Executor(handle, self, ctx, grad_req, group2ctx) executor.arg_arrays = args executor.grad_arrays = args_grad executor.aux_arrays = aux_states return executor
Gets the autodiff of current symbol. This function can only be used if current symbol is a loss function. .. note:: This function is currently not implemented. Parameters ---------- wrt : Array of String keyword arguments of the symbol that the gradients are taken. Returns ------- grad : Symbol A gradient Symbol with returns to be the corresponding gradients. def gradient(self, wrt): """Gets the autodiff of current symbol. This function can only be used if current symbol is a loss function. .. note:: This function is currently not implemented. Parameters ---------- wrt : Array of String keyword arguments of the symbol that the gradients are taken. Returns ------- grad : Symbol A gradient Symbol with returns to be the corresponding gradients. """ handle = SymbolHandle() c_wrt = c_str_array(wrt) check_call(_LIB.MXSymbolGrad(self.handle, mx_uint(len(wrt)), c_wrt, ctypes.byref(handle))) return Symbol(handle)
Evaluates a symbol given arguments. The `eval` method combines a call to `bind` (which returns an executor) with a call to `forward` (executor method). For the common use case, where you might repeatedly evaluate with same arguments, eval is slow. In that case, you should call `bind` once and then repeatedly call forward. This function allows simpler syntax for less cumbersome introspection. Example ------- >>> a = mx.sym.Variable('a') >>> b = mx.sym.Variable('b') >>> c = a + b >>> ex = c.eval(ctx = mx.cpu(), a = mx.nd.ones([2,3]), b = mx.nd.ones([2,3])) >>> ex [<NDArray 2x3 @cpu(0)>] >>> ex[0].asnumpy() array([[ 2., 2., 2.], [ 2., 2., 2.]], dtype=float32) Parameters ---------- ctx : Context The device context the generated executor to run on. kwargs : Keyword arguments of type `NDArray` Input arguments to the symbol. All the arguments must be provided. Returns ---------- result : a list of NDArrays corresponding to the values taken by each symbol when evaluated on given args. When called on a single symbol (not a group), the result will be a list with one element. def eval(self, ctx=None, **kwargs): """Evaluates a symbol given arguments. The `eval` method combines a call to `bind` (which returns an executor) with a call to `forward` (executor method). For the common use case, where you might repeatedly evaluate with same arguments, eval is slow. In that case, you should call `bind` once and then repeatedly call forward. This function allows simpler syntax for less cumbersome introspection. Example ------- >>> a = mx.sym.Variable('a') >>> b = mx.sym.Variable('b') >>> c = a + b >>> ex = c.eval(ctx = mx.cpu(), a = mx.nd.ones([2,3]), b = mx.nd.ones([2,3])) >>> ex [<NDArray 2x3 @cpu(0)>] >>> ex[0].asnumpy() array([[ 2., 2., 2.], [ 2., 2., 2.]], dtype=float32) Parameters ---------- ctx : Context The device context the generated executor to run on. kwargs : Keyword arguments of type `NDArray` Input arguments to the symbol. All the arguments must be provided. Returns ---------- result : a list of NDArrays corresponding to the values taken by each symbol when evaluated on given args. When called on a single symbol (not a group), the result will be a list with one element. """ if ctx is None: ctx = current_context() return self.bind(ctx, kwargs).forward()
Return symbol for target backend. Parameters ---------- backend : str The backend names. Returns ------- out : Symbol The created Symbol for target backend. def get_backend_symbol(self, backend): """Return symbol for target backend. Parameters ---------- backend : str The backend names. Returns ------- out : Symbol The created Symbol for target backend. """ out = SymbolHandle() check_call(_LIB.MXGenBackendSubgraph(self.handle, c_str(backend), ctypes.byref(out))) return Symbol(out)
Perform pixel-shuffling on the input. def hybrid_forward(self, F, x): """Perform pixel-shuffling on the input.""" f = self._factor # (N, C*f, W) x = F.reshape(x, (0, -4, -1, f, 0)) # (N, C, f, W) x = F.transpose(x, (0, 1, 3, 2)) # (N, C, W, f) x = F.reshape(x, (0, 0, -3)) # (N, C, W*f) return x
Perform pixel-shuffling on the input. def hybrid_forward(self, F, x): """Perform pixel-shuffling on the input.""" f1, f2 = self._factors # (N, f1*f2*C, H, W) x = F.reshape(x, (0, -4, -1, f1 * f2, 0, 0)) # (N, C, f1*f2, H, W) x = F.reshape(x, (0, 0, -4, f1, f2, 0, 0)) # (N, C, f1, f2, H, W) x = F.transpose(x, (0, 1, 4, 2, 5, 3)) # (N, C, H, f1, W, f2) x = F.reshape(x, (0, 0, -3, -3)) # (N, C, H*f1, W*f2) return x
Perform pixel-shuffling on the input. def hybrid_forward(self, F, x): """Perform pixel-shuffling on the input.""" # `transpose` doesn't support 8D, need other implementation f1, f2, f3 = self._factors # (N, C*f1*f2*f3, D, H, W) x = F.reshape(x, (0, -4, -1, f1 * f2 * f3, 0, 0, 0)) # (N, C, f1*f2*f3, D, H, W) x = F.swapaxes(x, 2, 3) # (N, C, D, f1*f2*f3, H, W) x = F.reshape(x, (0, 0, 0, -4, f1, f2*f3, 0, 0)) # (N, C, D, f1, f2*f3, H, W) x = F.reshape(x, (0, 0, -3, 0, 0, 0)) # (N, C, D*f1, f2*f3, H, W) x = F.swapaxes(x, 3, 4) # (N, C, D*f1, H, f2*f3, W) x = F.reshape(x, (0, 0, 0, 0, -4, f2, f3, 0)) # (N, C, D*f1, H, f2, f3, W) x = F.reshape(x, (0, 0, 0, -3, 0, 0)) # (N, C, D*f1, H*f2, f3, W) x = F.swapaxes(x, 4, 5) # (N, C, D*f1, H*f2, W, f3) x = F.reshape(x, (0, 0, 0, 0, -3)) # (N, C, D*f1, H*f2, W*f3) return x
Retry calling the decorated function using an exponential backoff. http://www.saltycrane.com/blog/2009/11/trying-out-retry-decorator-python/ original from: http://wiki.python.org/moin/PythonDecoratorLibrary#Retry :param target_exception: the exception to check. may be a tuple of exceptions to check :type target_exception: Exception or tuple :param tries: number of times to try (not retry) before giving up :type tries: int :param delay_s: initial delay between retries in seconds :type delay_s: int :param backoff: backoff multiplier e.g. value of 2 will double the delay each retry :type backoff: int def retry(target_exception, tries=4, delay_s=1, backoff=2): """Retry calling the decorated function using an exponential backoff. http://www.saltycrane.com/blog/2009/11/trying-out-retry-decorator-python/ original from: http://wiki.python.org/moin/PythonDecoratorLibrary#Retry :param target_exception: the exception to check. may be a tuple of exceptions to check :type target_exception: Exception or tuple :param tries: number of times to try (not retry) before giving up :type tries: int :param delay_s: initial delay between retries in seconds :type delay_s: int :param backoff: backoff multiplier e.g. value of 2 will double the delay each retry :type backoff: int """ import time from functools import wraps def decorated_retry(f): @wraps(f) def f_retry(*args, **kwargs): mtries, mdelay = tries, delay_s while mtries > 1: try: return f(*args, **kwargs) except target_exception as e: logging.warning("Exception: %s, Retrying in %d seconds...", str(e), mdelay) time.sleep(mdelay) mtries -= 1 mdelay *= backoff return f(*args, **kwargs) return f_retry # true decorator return decorated_retry
Returns a module loaded with the provided model. Parameters ---------- model_name: str Prefix of the MXNet model name as stored on the local directory. epoch_num : int Epoch number of model we would like to load. input_shape: tuple The shape of the input data in the form of (batch_size, channels, height, width) files: list of strings List of URLs pertaining to files that need to be downloaded in order to use the model. data_shapes: list of tuples. List of tuples where each tuple is a pair of input variable name and its shape. label_shapes: list of (str, tuple) Typically is ``data_iter.provide_label``. label_names: list of str Name of the output labels in the MXNet symbolic graph. gpus: str Comma separated string of gpu ids on which inferences are executed. E.g. 3,5,6 would refer to GPUs 3, 5 and 6. If empty, we use CPU. Returns ------- MXNet module def load_model(model_name, epoch_num, data_shapes, label_shapes, label_names, gpus=''): """Returns a module loaded with the provided model. Parameters ---------- model_name: str Prefix of the MXNet model name as stored on the local directory. epoch_num : int Epoch number of model we would like to load. input_shape: tuple The shape of the input data in the form of (batch_size, channels, height, width) files: list of strings List of URLs pertaining to files that need to be downloaded in order to use the model. data_shapes: list of tuples. List of tuples where each tuple is a pair of input variable name and its shape. label_shapes: list of (str, tuple) Typically is ``data_iter.provide_label``. label_names: list of str Name of the output labels in the MXNet symbolic graph. gpus: str Comma separated string of gpu ids on which inferences are executed. E.g. 3,5,6 would refer to GPUs 3, 5 and 6. If empty, we use CPU. Returns ------- MXNet module """ sym, arg_params, aux_params = mx.model.load_checkpoint(model_name, epoch_num) mod = create_module(sym, data_shapes, label_shapes, label_names, gpus) mod.set_params( arg_params=arg_params, aux_params=aux_params, allow_missing=True ) return mod
Creates a new MXNet module. Parameters ---------- sym : Symbol An MXNet symbol. input_shape: tuple The shape of the input data in the form of (batch_size, channels, height, width) files: list of strings List of URLs pertaining to files that need to be downloaded in order to use the model. data_shapes: list of tuples. List of tuples where each tuple is a pair of input variable name and its shape. label_shapes: list of (str, tuple) Typically is ``data_iter.provide_label``. label_names: list of str Name of the output labels in the MXNet symbolic graph. gpus: str Comma separated string of gpu ids on which inferences are executed. E.g. 3,5,6 would refer to GPUs 3, 5 and 6. If empty, we use CPU. Returns ------- MXNet module def create_module(sym, data_shapes, label_shapes, label_names, gpus=''): """Creates a new MXNet module. Parameters ---------- sym : Symbol An MXNet symbol. input_shape: tuple The shape of the input data in the form of (batch_size, channels, height, width) files: list of strings List of URLs pertaining to files that need to be downloaded in order to use the model. data_shapes: list of tuples. List of tuples where each tuple is a pair of input variable name and its shape. label_shapes: list of (str, tuple) Typically is ``data_iter.provide_label``. label_names: list of str Name of the output labels in the MXNet symbolic graph. gpus: str Comma separated string of gpu ids on which inferences are executed. E.g. 3,5,6 would refer to GPUs 3, 5 and 6. If empty, we use CPU. Returns ------- MXNet module """ if gpus == '': devices = mx.cpu() else: devices = [mx.gpu(int(i)) for i in gpus.split(',')] data_names = [data_shape[0] for data_shape in data_shapes] mod = mx.mod.Module( symbol=sym, data_names=data_names, context=devices, label_names=label_names ) mod.bind( for_training=False, data_shapes=data_shapes, label_shapes=label_shapes ) return mod
evalute network given validation record file Parameters: ---------- net : str or None Network name or use None to load from json without modifying path_imgrec : str path to the record validation file path_imglist : str path to the list file to replace labels in record file, optional num_classes : int number of classes, not including background mean_pixels : tuple (mean_r, mean_g, mean_b) data_shape : tuple or int (3, height, width) or height/width model_prefix : str model prefix of saved checkpoint epoch : int load model epoch ctx : mx.ctx mx.gpu() or mx.cpu() batch_size : int validation batch size nms_thresh : float non-maximum suppression threshold force_nms : boolean whether suppress different class objects ovp_thresh : float AP overlap threshold for true/false postives use_difficult : boolean whether to use difficult objects in evaluation if applicable class_names : comma separated str class names in string, must correspond to num_classes if set voc07_metric : boolean whether to use 11-point evluation as in VOC07 competition def evaluate_net(net, path_imgrec, num_classes, num_batch, mean_pixels, data_shape, model_prefix, epoch, ctx=mx.cpu(), batch_size=32, path_imglist="", nms_thresh=0.45, force_nms=False, ovp_thresh=0.5, use_difficult=False, class_names=None, voc07_metric=False): """ evalute network given validation record file Parameters: ---------- net : str or None Network name or use None to load from json without modifying path_imgrec : str path to the record validation file path_imglist : str path to the list file to replace labels in record file, optional num_classes : int number of classes, not including background mean_pixels : tuple (mean_r, mean_g, mean_b) data_shape : tuple or int (3, height, width) or height/width model_prefix : str model prefix of saved checkpoint epoch : int load model epoch ctx : mx.ctx mx.gpu() or mx.cpu() batch_size : int validation batch size nms_thresh : float non-maximum suppression threshold force_nms : boolean whether suppress different class objects ovp_thresh : float AP overlap threshold for true/false postives use_difficult : boolean whether to use difficult objects in evaluation if applicable class_names : comma separated str class names in string, must correspond to num_classes if set voc07_metric : boolean whether to use 11-point evluation as in VOC07 competition """ # set up logger logging.basicConfig() logger = logging.getLogger() logger.setLevel(logging.INFO) # args if isinstance(data_shape, int): data_shape = (3, data_shape, data_shape) assert len(data_shape) == 3 and data_shape[0] == 3 model_prefix += '_' + str(data_shape[1]) # iterator eval_iter = DetRecordIter(path_imgrec, batch_size, data_shape, mean_pixels=mean_pixels, path_imglist=path_imglist, **cfg.valid) # model params load_net, args, auxs = mx.model.load_checkpoint(model_prefix, epoch) # network if net is None: net = load_net else: net = get_symbol(net, data_shape[1], num_classes=num_classes, nms_thresh=nms_thresh, force_suppress=force_nms) if not 'label' in net.list_arguments(): label = mx.sym.Variable(name='label') net = mx.sym.Group([net, label]) # init module mod = mx.mod.Module(net, label_names=('label',), logger=logger, context=ctx, fixed_param_names=net.list_arguments()) mod.bind(data_shapes=eval_iter.provide_data, label_shapes=eval_iter.provide_label) mod.set_params(args, auxs, allow_missing=False, force_init=True) # run evaluation if voc07_metric: metric = VOC07MApMetric(ovp_thresh, use_difficult, class_names) else: metric = MApMetric(ovp_thresh, use_difficult, class_names) num = num_batch * batch_size data = [mx.random.uniform(-1.0, 1.0, shape=shape, ctx=ctx) for _, shape in mod.data_shapes] batch = mx.io.DataBatch(data, []) # empty label dry_run = 5 # use 5 iterations to warm up for i in range(dry_run): mod.forward(batch, is_train=False) for output in mod.get_outputs(): output.wait_to_read() tic = time.time() results = mod.score(eval_iter, metric, num_batch=num_batch) speed = num / (time.time() - tic) if logger is not None: logger.info('Finished inference with %d images' % num) logger.info('Finished with %f images per second', speed) for k, v in results: print("{}: {}".format(k, v))
Initializes the parameters and auxiliary states. By default this function does nothing. Subclass should override this method if contains parameters. Parameters ---------- initializer : Initializer Called to initialize parameters if needed. arg_params : dict If not ``None``, should be a dictionary of existing `arg_params`. Initialization will be copied from that. aux_params : dict If not ``None``, should be a dictionary of existing `aux_params`. Initialization will be copied from that. allow_missing : bool If ``True``, params could contain missing values, and the initializer will be called to fill those missing params. force_init : bool If ``True``, will force re-initialize even if already initialized. allow_extra : boolean, optional Whether allow extra parameters that are not needed by symbol. If this is True, no error will be thrown when arg_params or aux_params contain extra parameters that is not needed by the executor. def init_params(self, initializer=Uniform(0.01), arg_params=None, aux_params=None, allow_missing=False, force_init=False, allow_extra=False): """Initializes the parameters and auxiliary states. By default this function does nothing. Subclass should override this method if contains parameters. Parameters ---------- initializer : Initializer Called to initialize parameters if needed. arg_params : dict If not ``None``, should be a dictionary of existing `arg_params`. Initialization will be copied from that. aux_params : dict If not ``None``, should be a dictionary of existing `aux_params`. Initialization will be copied from that. allow_missing : bool If ``True``, params could contain missing values, and the initializer will be called to fill those missing params. force_init : bool If ``True``, will force re-initialize even if already initialized. allow_extra : boolean, optional Whether allow extra parameters that are not needed by symbol. If this is True, no error will be thrown when arg_params or aux_params contain extra parameters that is not needed by the executor. """ pass
Evaluates and accumulates evaluation metric on outputs of the last forward computation. Subclass should override this method if needed. Parameters ---------- eval_metric : EvalMetric labels : list of NDArray Typically ``data_batch.label``. def update_metric(self, eval_metric, labels, pre_sliced=False): """Evaluates and accumulates evaluation metric on outputs of the last forward computation. Subclass should override this method if needed. Parameters ---------- eval_metric : EvalMetric labels : list of NDArray Typically ``data_batch.label``. """ if self._label_shapes is None: # since we do not need labels, we are probably not a module with a loss # function or predictions, so just ignore this call return if pre_sliced: raise RuntimeError("PythonModule does not support presliced labels") # by default we expect our outputs are some scores that could be evaluated eval_metric.update(labels, self.get_outputs())
Binds the symbols to construct executors. This is necessary before one can perform computation with the module. Parameters ---------- data_shapes : list of (str, tuple) Typically is ``data_iter.provide_data``. label_shapes : list of (str, tuple) Typically is ``data_iter.provide_label``. for_training : bool Default is ``True``. Whether the executors should be bind for training. inputs_need_grad : bool Default is ``False``. Whether the gradients to the input data need to be computed. Typically this is not needed. But this might be needed when implementing composition of modules. force_rebind : bool Default is ``False``. This function does nothing if the executors are already bound. But with this ``True``, the executors will be forced to rebind. shared_module : Module Default is ``None``. This is used in bucketing. When not ``None``, the shared module essentially corresponds to a different bucket -- a module with different symbol but with the same sets of parameters (e.g. unrolled RNNs with different lengths). grad_req : str, list of str, dict of str to str Requirement for gradient accumulation. Can be 'write', 'add', or 'null' (default to 'write'). Can be specified globally (str) or for each argument (list, dict). def bind(self, data_shapes, label_shapes=None, for_training=True, inputs_need_grad=False, force_rebind=False, shared_module=None, grad_req='write'): """Binds the symbols to construct executors. This is necessary before one can perform computation with the module. Parameters ---------- data_shapes : list of (str, tuple) Typically is ``data_iter.provide_data``. label_shapes : list of (str, tuple) Typically is ``data_iter.provide_label``. for_training : bool Default is ``True``. Whether the executors should be bind for training. inputs_need_grad : bool Default is ``False``. Whether the gradients to the input data need to be computed. Typically this is not needed. But this might be needed when implementing composition of modules. force_rebind : bool Default is ``False``. This function does nothing if the executors are already bound. But with this ``True``, the executors will be forced to rebind. shared_module : Module Default is ``None``. This is used in bucketing. When not ``None``, the shared module essentially corresponds to a different bucket -- a module with different symbol but with the same sets of parameters (e.g. unrolled RNNs with different lengths). grad_req : str, list of str, dict of str to str Requirement for gradient accumulation. Can be 'write', 'add', or 'null' (default to 'write'). Can be specified globally (str) or for each argument (list, dict). """ if self.binded and not force_rebind: self.logger.warning('Already bound, ignoring bind()') return assert grad_req == 'write', "Python module only support write gradient" self.for_training = for_training self.inputs_need_grad = inputs_need_grad assert len(data_shapes) == len(self._data_names) assert [x[0] for x in data_shapes] == self._data_names self._data_shapes = data_shapes self._label_shapes = label_shapes if label_shapes is not None: assert self._label_names is not None assert len(self._label_names) == len(label_shapes) assert [x[0] for x in label_shapes] == self._label_names self._output_shapes = self._compute_output_shapes()
Forward computation. Here we do nothing but to keep a reference to the scores and the labels so that we can do backward computation. Parameters ---------- data_batch : DataBatch Could be anything with similar API implemented. is_train : bool Default is ``None``, which means `is_train` takes the value of ``self.for_training``. def forward(self, data_batch, is_train=None): """Forward computation. Here we do nothing but to keep a reference to the scores and the labels so that we can do backward computation. Parameters ---------- data_batch : DataBatch Could be anything with similar API implemented. is_train : bool Default is ``None``, which means `is_train` takes the value of ``self.for_training``. """ self._scores = data_batch.data[0] if is_train is None: is_train = self.for_training if is_train: self._labels = data_batch.label[0]
Actual implementation of the backward computation. The computation should take ``self._scores`` and ``self._labels`` and then compute the gradients with respect to the scores, store it as an `NDArray` in ``self._scores_grad``. Instead of defining a subclass and overriding this function, a more convenient way is to pass in a `grad_func` when constructing the module object. Then it will be called to compute the gradients. def _backward_impl(self): """Actual implementation of the backward computation. The computation should take ``self._scores`` and ``self._labels`` and then compute the gradients with respect to the scores, store it as an `NDArray` in ``self._scores_grad``. Instead of defining a subclass and overriding this function, a more convenient way is to pass in a `grad_func` when constructing the module object. Then it will be called to compute the gradients. """ if self._grad_func is not None: grad = self._grad_func(self._scores, self._labels) if not isinstance(grad, nd.NDArray): grad = nd.array(grad) self._scores_grad = grad else: raise NotImplementedError()
Encode sentences and (optionally) build a mapping from string tokens to integer indices. Unknown keys will be added to vocabulary. Parameters ---------- sentences : list of list of str A list of sentences to encode. Each sentence should be a list of string tokens. vocab : None or dict of str -> int Optional input Vocabulary invalid_label : int, default -1 Index for invalid token, like <end-of-sentence> invalid_key : str, default '\\n' Key for invalid token. Use '\\n' for end of sentence by default. start_label : int lowest index. unknown_token: str Symbol to represent unknown token. If not specified, unknown token will be skipped. Returns ------- result : list of list of int encoded sentences vocab : dict of str -> int result vocabulary def encode_sentences(sentences, vocab=None, invalid_label=-1, invalid_key='\n', start_label=0, unknown_token=None): """Encode sentences and (optionally) build a mapping from string tokens to integer indices. Unknown keys will be added to vocabulary. Parameters ---------- sentences : list of list of str A list of sentences to encode. Each sentence should be a list of string tokens. vocab : None or dict of str -> int Optional input Vocabulary invalid_label : int, default -1 Index for invalid token, like <end-of-sentence> invalid_key : str, default '\\n' Key for invalid token. Use '\\n' for end of sentence by default. start_label : int lowest index. unknown_token: str Symbol to represent unknown token. If not specified, unknown token will be skipped. Returns ------- result : list of list of int encoded sentences vocab : dict of str -> int result vocabulary """ idx = start_label if vocab is None: vocab = {invalid_key: invalid_label} new_vocab = True else: new_vocab = False res = [] for sent in sentences: coded = [] for word in sent: if word not in vocab: assert (new_vocab or unknown_token), "Unknown token %s"%word if idx == invalid_label: idx += 1 if unknown_token: word = unknown_token vocab[word] = idx idx += 1 coded.append(vocab[word]) res.append(coded) return res, vocab
Resets the iterator to the beginning of the data. def reset(self): """Resets the iterator to the beginning of the data.""" self.curr_idx = 0 random.shuffle(self.idx) for buck in self.data: np.random.shuffle(buck) self.nddata = [] self.ndlabel = [] for buck in self.data: label = np.empty_like(buck) label[:, :-1] = buck[:, 1:] label[:, -1] = self.invalid_label self.nddata.append(ndarray.array(buck, dtype=self.dtype)) self.ndlabel.append(ndarray.array(label, dtype=self.dtype))
Returns the next batch of data. def next(self): """Returns the next batch of data.""" if self.curr_idx == len(self.idx): raise StopIteration i, j = self.idx[self.curr_idx] self.curr_idx += 1 if self.major_axis == 1: data = self.nddata[i][j:j+self.batch_size].T label = self.ndlabel[i][j:j+self.batch_size].T else: data = self.nddata[i][j:j+self.batch_size] label = self.ndlabel[i][j:j+self.batch_size] return DataBatch([data], [label], pad=0, bucket_key=self.buckets[i], provide_data=[DataDesc( name=self.data_name, shape=data.shape, layout=self.layout)], provide_label=[DataDesc( name=self.label_name, shape=label.shape, layout=self.layout)])
Returns the singleton instance. Upon its first call, it creates a new instance of the decorated class and calls its `__init__` method. On all subsequent calls, the already created instance is returned. def getInstance(self): """ Returns the singleton instance. Upon its first call, it creates a new instance of the decorated class and calls its `__init__` method. On all subsequent calls, the already created instance is returned. """ try: return self._instance except AttributeError: self._instance = self._decorated() return self._instance
Description : run lipnet training code using argument info def main(): """ Description : run lipnet training code using argument info """ parser = argparse.ArgumentParser() parser.add_argument('--batch_size', type=int, default=64) parser.add_argument('--image_path', type=str, default='./data/datasets/') parser.add_argument('--align_path', type=str, default='./data/align/') parser.add_argument('--num_gpus', type=int, default=1) parser.add_argument('--num_workers', type=int, default=0) parser.add_argument('--data_type', type=str, default='valid') parser.add_argument('--model_path', type=str, default=None) config = parser.parse_args() trainer = Train(config) trainer.build_model(path=config.model_path) trainer.load_dataloader() if config.data_type == 'train': data_loader = trainer.train_dataloader elif config.data_type == 'valid': data_loader = trainer.valid_dataloader trainer.infer_batch(data_loader)
Get the variable given a name if one exists or create a new one if missing. Parameters ---------- name : str name of the variable **kwargs : more arguments that's passed to symbol.Variable def get(self, name, **kwargs): """Get the variable given a name if one exists or create a new one if missing. Parameters ---------- name : str name of the variable **kwargs : more arguments that's passed to symbol.Variable """ name = self._prefix + name if name not in self._params: self._params[name] = symbol.Variable(name, **kwargs) return self._params[name]
Reset before re-using the cell for another graph. def reset(self): """Reset before re-using the cell for another graph.""" self._init_counter = -1 self._counter = -1 if hasattr(self, '_cells'): for cell in self._cells: cell.reset()
Initial state for this cell. Parameters ---------- func : callable, default symbol.zeros Function for creating initial state. Can be symbol.zeros, symbol.uniform, symbol.Variable etc. Use symbol.Variable if you want to directly feed input as states. **kwargs : more keyword arguments passed to func. For example mean, std, dtype, etc. Returns ------- states : nested list of Symbol Starting states for the first RNN step. def begin_state(self, func=symbol.zeros, **kwargs): """Initial state for this cell. Parameters ---------- func : callable, default symbol.zeros Function for creating initial state. Can be symbol.zeros, symbol.uniform, symbol.Variable etc. Use symbol.Variable if you want to directly feed input as states. **kwargs : more keyword arguments passed to func. For example mean, std, dtype, etc. Returns ------- states : nested list of Symbol Starting states for the first RNN step. """ assert not self._modified, \ "After applying modifier cells (e.g. DropoutCell) the base " \ "cell cannot be called directly. Call the modifier cell instead." states = [] for info in self.state_info: self._init_counter += 1 if info is None: state = func(name='%sbegin_state_%d'%(self._prefix, self._init_counter), **kwargs) else: kwargs.update(info) state = func(name='%sbegin_state_%d'%(self._prefix, self._init_counter), **kwargs) states.append(state) return states
Unpack fused weight matrices into separate weight matrices. For example, say you use a module object `mod` to run a network that has an lstm cell. In `mod.get_params()[0]`, the lstm parameters are all represented as a single big vector. `cell.unpack_weights(mod.get_params()[0])` will unpack this vector into a dictionary of more readable lstm parameters - c, f, i, o gates for i2h (input to hidden) and h2h (hidden to hidden) weights. Parameters ---------- args : dict of str -> NDArray Dictionary containing packed weights. usually from `Module.get_params()[0]`. Returns ------- args : dict of str -> NDArray Dictionary with unpacked weights associated with this cell. See Also -------- pack_weights: Performs the reverse operation of this function. def unpack_weights(self, args): """Unpack fused weight matrices into separate weight matrices. For example, say you use a module object `mod` to run a network that has an lstm cell. In `mod.get_params()[0]`, the lstm parameters are all represented as a single big vector. `cell.unpack_weights(mod.get_params()[0])` will unpack this vector into a dictionary of more readable lstm parameters - c, f, i, o gates for i2h (input to hidden) and h2h (hidden to hidden) weights. Parameters ---------- args : dict of str -> NDArray Dictionary containing packed weights. usually from `Module.get_params()[0]`. Returns ------- args : dict of str -> NDArray Dictionary with unpacked weights associated with this cell. See Also -------- pack_weights: Performs the reverse operation of this function. """ args = args.copy() if not self._gate_names: return args h = self._num_hidden for group_name in ['i2h', 'h2h']: weight = args.pop('%s%s_weight'%(self._prefix, group_name)) bias = args.pop('%s%s_bias' % (self._prefix, group_name)) for j, gate in enumerate(self._gate_names): wname = '%s%s%s_weight' % (self._prefix, group_name, gate) args[wname] = weight[j*h:(j+1)*h].copy() bname = '%s%s%s_bias' % (self._prefix, group_name, gate) args[bname] = bias[j*h:(j+1)*h].copy() return args
Pack separate weight matrices into a single packed weight. Parameters ---------- args : dict of str -> NDArray Dictionary containing unpacked weights. Returns ------- args : dict of str -> NDArray Dictionary with packed weights associated with this cell. def pack_weights(self, args): """Pack separate weight matrices into a single packed weight. Parameters ---------- args : dict of str -> NDArray Dictionary containing unpacked weights. Returns ------- args : dict of str -> NDArray Dictionary with packed weights associated with this cell. """ args = args.copy() if not self._gate_names: return args for group_name in ['i2h', 'h2h']: weight = [] bias = [] for gate in self._gate_names: wname = '%s%s%s_weight'%(self._prefix, group_name, gate) weight.append(args.pop(wname)) bname = '%s%s%s_bias'%(self._prefix, group_name, gate) bias.append(args.pop(bname)) args['%s%s_weight'%(self._prefix, group_name)] = ndarray.concatenate(weight) args['%s%s_bias'%(self._prefix, group_name)] = ndarray.concatenate(bias) return args
Unroll an RNN cell across time steps. Parameters ---------- length : int Number of steps to unroll. inputs : Symbol, list of Symbol, or None If `inputs` is a single Symbol (usually the output of Embedding symbol), it should have shape (batch_size, length, ...) if layout == 'NTC', or (length, batch_size, ...) if layout == 'TNC'. If `inputs` is a list of symbols (usually output of previous unroll), they should all have shape (batch_size, ...). begin_state : nested list of Symbol, default None Input states created by `begin_state()` or output state of another cell. Created from `begin_state()` if None. layout : str, optional `layout` of input symbol. Only used if inputs is a single Symbol. merge_outputs : bool, optional If False, return outputs as a list of Symbols. If True, concatenate output across time steps and return a single symbol with shape (batch_size, length, ...) if layout == 'NTC', or (length, batch_size, ...) if layout == 'TNC'. If None, output whatever is faster. Returns ------- outputs : list of Symbol or Symbol Symbol (if `merge_outputs` is True) or list of Symbols (if `merge_outputs` is False) corresponding to the output from the RNN from this unrolling. states : nested list of Symbol The new state of this RNN after this unrolling. The type of this symbol is same as the output of begin_state(). def unroll(self, length, inputs, begin_state=None, layout='NTC', merge_outputs=None): """Unroll an RNN cell across time steps. Parameters ---------- length : int Number of steps to unroll. inputs : Symbol, list of Symbol, or None If `inputs` is a single Symbol (usually the output of Embedding symbol), it should have shape (batch_size, length, ...) if layout == 'NTC', or (length, batch_size, ...) if layout == 'TNC'. If `inputs` is a list of symbols (usually output of previous unroll), they should all have shape (batch_size, ...). begin_state : nested list of Symbol, default None Input states created by `begin_state()` or output state of another cell. Created from `begin_state()` if None. layout : str, optional `layout` of input symbol. Only used if inputs is a single Symbol. merge_outputs : bool, optional If False, return outputs as a list of Symbols. If True, concatenate output across time steps and return a single symbol with shape (batch_size, length, ...) if layout == 'NTC', or (length, batch_size, ...) if layout == 'TNC'. If None, output whatever is faster. Returns ------- outputs : list of Symbol or Symbol Symbol (if `merge_outputs` is True) or list of Symbols (if `merge_outputs` is False) corresponding to the output from the RNN from this unrolling. states : nested list of Symbol The new state of this RNN after this unrolling. The type of this symbol is same as the output of begin_state(). """ self.reset() inputs, _ = _normalize_sequence(length, inputs, layout, False) if begin_state is None: begin_state = self.begin_state() states = begin_state outputs = [] for i in range(length): output, states = self(inputs[i], states) outputs.append(output) outputs, _ = _normalize_sequence(length, outputs, layout, merge_outputs) return outputs, states
Get activation function. Convert if is string def _get_activation(self, inputs, activation, **kwargs): """Get activation function. Convert if is string""" if isinstance(activation, string_types): return symbol.Activation(inputs, act_type=activation, **kwargs) else: return activation(inputs, **kwargs)
slice fused rnn weights def _slice_weights(self, arr, li, lh): """slice fused rnn weights""" args = {} gate_names = self._gate_names directions = self._directions b = len(directions) p = 0 for layer in range(self._num_layers): for direction in directions: for gate in gate_names: name = '%s%s%d_i2h%s_weight'%(self._prefix, direction, layer, gate) if layer > 0: size = b*lh*lh args[name] = arr[p:p+size].reshape((lh, b*lh)) else: size = li*lh args[name] = arr[p:p+size].reshape((lh, li)) p += size for gate in gate_names: name = '%s%s%d_h2h%s_weight'%(self._prefix, direction, layer, gate) size = lh**2 args[name] = arr[p:p+size].reshape((lh, lh)) p += size for layer in range(self._num_layers): for direction in directions: for gate in gate_names: name = '%s%s%d_i2h%s_bias'%(self._prefix, direction, layer, gate) args[name] = arr[p:p+lh] p += lh for gate in gate_names: name = '%s%s%d_h2h%s_bias'%(self._prefix, direction, layer, gate) args[name] = arr[p:p+lh] p += lh assert p == arr.size, "Invalid parameters size for FusedRNNCell" return args
Unfuse the fused RNN in to a stack of rnn cells. Returns ------- cell : mxnet.rnn.SequentialRNNCell unfused cell that can be used for stepping, and can run on CPU. def unfuse(self): """Unfuse the fused RNN in to a stack of rnn cells. Returns ------- cell : mxnet.rnn.SequentialRNNCell unfused cell that can be used for stepping, and can run on CPU. """ stack = SequentialRNNCell() get_cell = {'rnn_relu': lambda cell_prefix: RNNCell(self._num_hidden, activation='relu', prefix=cell_prefix), 'rnn_tanh': lambda cell_prefix: RNNCell(self._num_hidden, activation='tanh', prefix=cell_prefix), 'lstm': lambda cell_prefix: LSTMCell(self._num_hidden, prefix=cell_prefix), 'gru': lambda cell_prefix: GRUCell(self._num_hidden, prefix=cell_prefix)}[self._mode] for i in range(self._num_layers): if self._bidirectional: stack.add(BidirectionalCell( get_cell('%sl%d_'%(self._prefix, i)), get_cell('%sr%d_'%(self._prefix, i)), output_prefix='%sbi_l%d_'%(self._prefix, i))) else: stack.add(get_cell('%sl%d_'%(self._prefix, i))) if self._dropout > 0 and i != self._num_layers - 1: stack.add(DropoutCell(self._dropout, prefix='%s_dropout%d_'%(self._prefix, i))) return stack
Append a cell into the stack. Parameters ---------- cell : BaseRNNCell The cell to be appended. During unroll, previous cell's output (or raw inputs if no previous cell) is used as the input to this cell. def add(self, cell): """Append a cell into the stack. Parameters ---------- cell : BaseRNNCell The cell to be appended. During unroll, previous cell's output (or raw inputs if no previous cell) is used as the input to this cell. """ self._cells.append(cell) if self._override_cell_params: assert cell._own_params, \ "Either specify params for SequentialRNNCell " \ "or child cells, not both." cell.params._params.update(self.params._params) self.params._params.update(cell.params._params)
Reads an image from file path or URL, optionally resizing to given image dimensions and subtracting mean. :param img_path: path to file, or url to download :param image_dims: image dimensions to resize to, or None :param mean: mean file to subtract, or None :return: loaded image, in RGB format def read_image(img_path, image_dims=None, mean=None): """ Reads an image from file path or URL, optionally resizing to given image dimensions and subtracting mean. :param img_path: path to file, or url to download :param image_dims: image dimensions to resize to, or None :param mean: mean file to subtract, or None :return: loaded image, in RGB format """ import urllib filename = img_path.split("/")[-1] if img_path.startswith('http'): urllib.urlretrieve(img_path, filename) img = cv2.imread(filename) else: img = cv2.imread(img_path) img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) if image_dims is not None: img = cv2.resize(img, image_dims) # resize to image_dims to fit model img = np.rollaxis(img, 2) # change to (c, h, w) order img = img[np.newaxis, :] # extend to (n, c, h, w) if mean is not None: mean = np.array(mean) if mean.shape == (3,): mean = mean[np.newaxis, :, np.newaxis, np.newaxis] # extend to (n, c, 1, 1) img = img.astype(np.float32) - mean # subtract mean return img
Changes device of given mxnet arguments :param arg_params: arguments :param aux_params: auxiliary parameters :param ctx: new device context :return: arguments and auxiliary parameters on new device def _ch_dev(arg_params, aux_params, ctx): """ Changes device of given mxnet arguments :param arg_params: arguments :param aux_params: auxiliary parameters :param ctx: new device context :return: arguments and auxiliary parameters on new device """ new_args = dict() new_auxs = dict() for k, v in arg_params.items(): new_args[k] = v.as_in_context(ctx) for k, v in aux_params.items(): new_auxs[k] = v.as_in_context(ctx) return new_args, new_auxs
Run the layer comparison on a caffe model, given its prototxt, weights and mean. The comparison is done by inferring on a given image using both caffe and mxnet model :param image_url: image file or url to run inference on :param gpu: gpu to use, -1 for cpu :param caffe_prototxt_path: path to caffe prototxt :param caffe_model_path: path to caffe weights :param caffe_mean: path to caffe mean file def convert_and_compare_caffe_to_mxnet(image_url, gpu, caffe_prototxt_path, caffe_model_path, caffe_mean, mean_diff_allowed, max_diff_allowed): """ Run the layer comparison on a caffe model, given its prototxt, weights and mean. The comparison is done by inferring on a given image using both caffe and mxnet model :param image_url: image file or url to run inference on :param gpu: gpu to use, -1 for cpu :param caffe_prototxt_path: path to caffe prototxt :param caffe_model_path: path to caffe weights :param caffe_mean: path to caffe mean file """ import caffe from caffe_proto_utils import read_network_dag, process_network_proto, read_caffe_mean from convert_model import convert_model if isinstance(caffe_mean, str): caffe_mean = read_caffe_mean(caffe_mean) elif caffe_mean is None: pass elif len(caffe_mean) == 3: # swap channels from Caffe BGR to RGB caffe_mean = caffe_mean[::-1] # get caffe root location, this is needed to run the upgrade network utility, so we only need # to support parsing of latest caffe caffe_root = os.path.dirname(os.path.dirname(caffe.__path__[0])) caffe_prototxt_path = process_network_proto(caffe_root, caffe_prototxt_path) _, layer_name_to_record, top_to_layers = read_network_dag(caffe_prototxt_path) caffe.set_mode_cpu() caffe_net = caffe.Net(caffe_prototxt_path, caffe_model_path, caffe.TEST) image_dims = tuple(caffe_net.blobs['data'].shape)[2:4] logging.info('getting image %s', image_url) img_rgb = read_image(image_url, image_dims, caffe_mean) img_bgr = img_rgb[:, ::-1, :, :] caffe_net.blobs['data'].reshape(*img_bgr.shape) caffe_net.blobs['data'].data[...] = img_bgr _ = caffe_net.forward() # read sym and add all outputs sym, arg_params, aux_params, _ = convert_model(caffe_prototxt_path, caffe_model_path) sym = sym.get_internals() # now mxnet if gpu < 0: ctx = mx.cpu(0) else: ctx = mx.gpu(gpu) arg_params, aux_params = _ch_dev(arg_params, aux_params, ctx) arg_params["data"] = mx.nd.array(img_rgb, ctx) arg_params["prob_label"] = mx.nd.empty((1,), ctx) exe = sym.bind(ctx, arg_params, args_grad=None, grad_req="null", aux_states=aux_params) exe.forward(is_train=False) compare_layers_from_nets(caffe_net, arg_params, aux_params, exe, layer_name_to_record, top_to_layers, mean_diff_allowed, max_diff_allowed) return
Implementation of Breadth-first search (BFS) on caffe network DAG :param root_node: root node of caffe network DAG :param process_node: function to run on each node def _bfs(root_node, process_node): """ Implementation of Breadth-first search (BFS) on caffe network DAG :param root_node: root node of caffe network DAG :param process_node: function to run on each node """ from collections import deque seen_nodes = set() next_nodes = deque() seen_nodes.add(root_node) next_nodes.append(root_node) while next_nodes: current_node = next_nodes.popleft() # process current node process_node(current_node) for child_node in current_node.children: if child_node not in seen_nodes: seen_nodes.add(child_node) next_nodes.append(child_node)
Compare layer by layer of a caffe network with mxnet network :param caffe_net: loaded caffe network :param arg_params: arguments :param aux_params: auxiliary parameters :param exe: mxnet model :param layer_name_to_record: map between caffe layer and information record :param top_to_layers: map between caffe blob name to layers which outputs it (including inplace) :param mean_diff_allowed: mean difference allowed between caffe blob and mxnet blob :param max_diff_allowed: max difference allowed between caffe blob and mxnet blob def compare_layers_from_nets(caffe_net, arg_params, aux_params, exe, layer_name_to_record, top_to_layers, mean_diff_allowed, max_diff_allowed): """ Compare layer by layer of a caffe network with mxnet network :param caffe_net: loaded caffe network :param arg_params: arguments :param aux_params: auxiliary parameters :param exe: mxnet model :param layer_name_to_record: map between caffe layer and information record :param top_to_layers: map between caffe blob name to layers which outputs it (including inplace) :param mean_diff_allowed: mean difference allowed between caffe blob and mxnet blob :param max_diff_allowed: max difference allowed between caffe blob and mxnet blob """ import re log_format = ' {0:<40} {1:<40} {2:<8} {3:>10} {4:>10} {5:<1}' compare_layers_from_nets.is_first_convolution = True def _compare_blob(caf_blob, mx_blob, caf_name, mx_name, blob_type, note): diff = np.abs(mx_blob - caf_blob) diff_mean = diff.mean() diff_max = diff.max() logging.info(log_format.format(caf_name, mx_name, blob_type, '%4.5f' % diff_mean, '%4.5f' % diff_max, note)) assert diff_mean < mean_diff_allowed assert diff_max < max_diff_allowed def _process_layer_parameters(layer): logging.debug('processing layer %s of type %s', layer.name, layer.type) normalized_layer_name = re.sub('[-/]', '_', layer.name) # handle weight and bias of convolution and fully-connected layers if layer.name in caffe_net.params and layer.type in ['Convolution', 'InnerProduct', 'Deconvolution']: has_bias = len(caffe_net.params[layer.name]) > 1 mx_name_weight = '{}_weight'.format(normalized_layer_name) mx_beta = arg_params[mx_name_weight].asnumpy() # first convolution should change from BGR to RGB if layer.type == 'Convolution' and compare_layers_from_nets.is_first_convolution: compare_layers_from_nets.is_first_convolution = False # if RGB or RGBA if mx_beta.shape[1] == 3 or mx_beta.shape[1] == 4: # Swapping BGR of caffe into RGB in mxnet mx_beta[:, [0, 2], :, :] = mx_beta[:, [2, 0], :, :] caf_beta = caffe_net.params[layer.name][0].data _compare_blob(caf_beta, mx_beta, layer.name, mx_name_weight, 'weight', '') if has_bias: mx_name_bias = '{}_bias'.format(normalized_layer_name) mx_gamma = arg_params[mx_name_bias].asnumpy() caf_gamma = caffe_net.params[layer.name][1].data _compare_blob(caf_gamma, mx_gamma, layer.name, mx_name_bias, 'bias', '') elif layer.name in caffe_net.params and layer.type == 'Scale': if 'scale' in normalized_layer_name: bn_name = normalized_layer_name.replace('scale', 'bn') elif 'sc' in normalized_layer_name: bn_name = normalized_layer_name.replace('sc', 'bn') else: assert False, 'Unknown name convention for bn/scale' beta_name = '{}_beta'.format(bn_name) gamma_name = '{}_gamma'.format(bn_name) mx_beta = arg_params[beta_name].asnumpy() caf_beta = caffe_net.params[layer.name][1].data _compare_blob(caf_beta, mx_beta, layer.name, beta_name, 'mov_mean', '') mx_gamma = arg_params[gamma_name].asnumpy() caf_gamma = caffe_net.params[layer.name][0].data _compare_blob(caf_gamma, mx_gamma, layer.name, gamma_name, 'mov_var', '') elif layer.name in caffe_net.params and layer.type == 'BatchNorm': mean_name = '{}_moving_mean'.format(normalized_layer_name) var_name = '{}_moving_var'.format(normalized_layer_name) caf_rescale_factor = caffe_net.params[layer.name][2].data mx_mean = aux_params[mean_name].asnumpy() caf_mean = caffe_net.params[layer.name][0].data / caf_rescale_factor _compare_blob(caf_mean, mx_mean, layer.name, mean_name, 'mean', '') mx_var = aux_params[var_name].asnumpy() caf_var = caffe_net.params[layer.name][1].data / caf_rescale_factor _compare_blob(caf_var, mx_var, layer.name, var_name, 'var', 'expect 1e-04 change due to cudnn eps') elif layer.type in ['Input', 'Pooling', 'ReLU', 'Eltwise', 'Softmax', 'LRN', 'Concat', 'Dropout', 'Crop']: # no parameters to check for these layers pass else: warnings.warn('No handling for layer %s of type %s, should we ignore it?', layer.name, layer.type) return def _process_layer_output(caffe_blob_name): logging.debug('processing blob %s', caffe_blob_name) # skip blobs not originating from actual layers, e.g. artificial split layers added by caffe if caffe_blob_name not in top_to_layers: return caf_blob = caffe_net.blobs[caffe_blob_name].data # data should change from BGR to RGB if caffe_blob_name == 'data': # if RGB or RGBA if caf_blob.shape[1] == 3 or caf_blob.shape[1] == 4: # Swapping BGR of caffe into RGB in mxnet caf_blob[:, [0, 2], :, :] = caf_blob[:, [2, 0], :, :] mx_name = 'data' else: # get last layer name which outputs this blob name last_layer_name = top_to_layers[caffe_blob_name][-1] normalized_last_layer_name = re.sub('[-/]', '_', last_layer_name) mx_name = '{}_output'.format(normalized_last_layer_name) if 'scale' in mx_name: mx_name = mx_name.replace('scale', 'bn') elif 'sc' in mx_name: mx_name = mx_name.replace('sc', 'bn') if mx_name not in exe.output_dict: logging.error('mxnet blob %s is missing, time to extend the compare tool..', mx_name) return mx_blob = exe.output_dict[mx_name].asnumpy() _compare_blob(caf_blob, mx_blob, caffe_blob_name, mx_name, 'output', '') return # check layer parameters logging.info('\n***** Network Parameters '.ljust(140, '*')) logging.info(log_format.format('CAFFE', 'MXNET', 'Type', 'Mean(diff)', 'Max(diff)', 'Note')) first_layer_name = layer_name_to_record.keys()[0] _bfs(layer_name_to_record[first_layer_name], _process_layer_parameters) # check layer output logging.info('\n***** Network Outputs '.ljust(140, '*')) logging.info(log_format.format('CAFFE', 'MXNET', 'Type', 'Mean(diff)', 'Max(diff)', 'Note')) for caffe_blob_name in caffe_net.blobs.keys(): _process_layer_output(caffe_blob_name) return
Entrypoint for compare_layers def main(): """Entrypoint for compare_layers""" parser = argparse.ArgumentParser( description='Tool for testing caffe to mxnet conversion layer by layer') parser.add_argument('--image_url', type=str, default='https://github.com/dmlc/web-data/raw/master/mxnet/doc/'\ 'tutorials/python/predict_image/cat.jpg', help='input image to test inference, can be either file path or url') parser.add_argument('--caffe_prototxt_path', type=str, default='./model.prototxt', help='path to caffe prototxt') parser.add_argument('--caffe_model_path', type=str, default='./model.caffemodel', help='path to caffe weights') parser.add_argument('--caffe_mean', type=str, default='./model_mean.binaryproto', help='path to caffe mean file') parser.add_argument('--mean_diff_allowed', type=int, default=1e-03, help='mean difference allowed between caffe blob and mxnet blob') parser.add_argument('--max_diff_allowed', type=int, default=1e-01, help='max difference allowed between caffe blob and mxnet blob') parser.add_argument('--gpu', type=int, default=-1, help='the gpu id used for predict') args = parser.parse_args() convert_and_compare_caffe_to_mxnet(args.image_url, args.gpu, args.caffe_prototxt_path, args.caffe_model_path, args.caffe_mean, args.mean_diff_allowed, args.max_diff_allowed)
Get executor to Stochastic Gradient Langevin Dynamics and/or Bayesian Dark Knowledge def get_executor(sym, ctx, data_inputs, initializer=None): """Get executor to Stochastic Gradient Langevin Dynamics and/or Bayesian Dark Knowledge""" data_shapes = {k: v.shape for k, v in data_inputs.items()} arg_names = sym.list_arguments() aux_names = sym.list_auxiliary_states() param_names = list(set(arg_names) - set(data_inputs.keys())) arg_shapes, output_shapes, aux_shapes = sym.infer_shape(**data_shapes) arg_name_shape = {k: s for k, s in zip(arg_names, arg_shapes)} params = {n: nd.empty(arg_name_shape[n], ctx=ctx) for n in param_names} params_grad = {n: nd.empty(arg_name_shape[n], ctx=ctx) for n in param_names} aux_states = {k: nd.empty(s, ctx=ctx) for k, s in zip(aux_names, aux_shapes)} exe = sym.bind(ctx=ctx, args=dict(params, **data_inputs), args_grad=params_grad, aux_states=aux_states) if initializer is not None: for k, v in params.items(): initializer(k, v) return exe, params, params_grad, aux_states
Create copy of parameters def copy_param(exe, new_param=None): """Create copy of parameters""" if new_param is None: new_param = {k: nd.empty(v.shape, ctx=mx.cpu()) for k, v in exe.arg_dict.items()} for k, v in new_param.items(): exe.arg_dict[k].copyto(v) return new_param
Parse command line arguments def parse_args(): """Parse command line arguments""" parser = argparse.ArgumentParser() parser.add_argument("font_path", help="Path to ttf font file or directory containing ttf files") parser.add_argument("--loss", help="'ctc' or 'warpctc' loss [Default 'ctc']", default='ctc') parser.add_argument("--cpu", help="Number of CPUs for training [Default 8]. Ignored if --gpu is specified.", type=int, default=8) parser.add_argument("--gpu", help="Number of GPUs for training [Default 0]", type=int) parser.add_argument("--num_proc", help="Number CAPTCHA generating processes [Default 4]", type=int, default=4) parser.add_argument("--prefix", help="Checkpoint prefix [Default 'ocr']", default='ocr') return parser.parse_args()
Program entry point def main(): """Program entry point""" args = parse_args() if not any(args.loss == s for s in ['ctc', 'warpctc']): raise ValueError("Invalid loss '{}' (must be 'ctc' or 'warpctc')".format(args.loss)) hp = Hyperparams() # Start a multiprocessor captcha image generator mp_captcha = MPDigitCaptcha( font_paths=get_fonts(args.font_path), h=hp.seq_length, w=30, num_digit_min=3, num_digit_max=4, num_processes=args.num_proc, max_queue_size=hp.batch_size * 2) try: # Must call start() before any call to mxnet module (https://github.com/apache/incubator-mxnet/issues/9213) mp_captcha.start() if args.gpu: contexts = [mx.context.gpu(i) for i in range(args.gpu)] else: contexts = [mx.context.cpu(i) for i in range(args.cpu)] init_states = lstm.init_states(hp.batch_size, hp.num_lstm_layer, hp.num_hidden) data_train = OCRIter( hp.train_epoch_size // hp.batch_size, hp.batch_size, init_states, captcha=mp_captcha, name='train') data_val = OCRIter( hp.eval_epoch_size // hp.batch_size, hp.batch_size, init_states, captcha=mp_captcha, name='val') symbol = lstm.lstm_unroll( num_lstm_layer=hp.num_lstm_layer, seq_len=hp.seq_length, num_hidden=hp.num_hidden, num_label=hp.num_label, loss_type=args.loss) head = '%(asctime)-15s %(message)s' logging.basicConfig(level=logging.DEBUG, format=head) module = mx.mod.Module( symbol, data_names=['data', 'l0_init_c', 'l0_init_h', 'l1_init_c', 'l1_init_h'], label_names=['label'], context=contexts) metrics = CtcMetrics(hp.seq_length) module.fit(train_data=data_train, eval_data=data_val, # use metrics.accuracy or metrics.accuracy_lcs eval_metric=mx.metric.np(metrics.accuracy, allow_extra_outputs=True), optimizer='sgd', optimizer_params={'learning_rate': hp.learning_rate, 'momentum': hp.momentum, 'wd': 0.00001, }, initializer=mx.init.Xavier(factor_type="in", magnitude=2.34), num_epoch=hp.num_epoch, batch_end_callback=mx.callback.Speedometer(hp.batch_size, 50), epoch_end_callback=mx.callback.do_checkpoint(args.prefix), ) except KeyboardInterrupt: print("W: interrupt received, stopping...") finally: # Reset multiprocessing captcha generator to stop processes mp_captcha.reset()
Gatys et al. CVPR 2017 ref: Image Style Transfer Using Convolutional Neural Networks def optimize(args): """ Gatys et al. CVPR 2017 ref: Image Style Transfer Using Convolutional Neural Networks """ if args.cuda: ctx = mx.gpu(0) else: ctx = mx.cpu(0) # load the content and style target content_image = utils.tensor_load_rgbimage(args.content_image,ctx, size=args.content_size, keep_asp=True) content_image = utils.subtract_imagenet_mean_preprocess_batch(content_image) style_image = utils.tensor_load_rgbimage(args.style_image, ctx, size=args.style_size) style_image = utils.subtract_imagenet_mean_preprocess_batch(style_image) # load the pre-trained vgg-16 and extract features vgg = net.Vgg16() utils.init_vgg_params(vgg, 'models', ctx=ctx) # content feature f_xc_c = vgg(content_image)[1] # style feature features_style = vgg(style_image) gram_style = [net.gram_matrix(y) for y in features_style] # output output = Parameter('output', shape=content_image.shape) output.initialize(ctx=ctx) output.set_data(content_image) # optimizer trainer = gluon.Trainer([output], 'adam', {'learning_rate': args.lr}) mse_loss = gluon.loss.L2Loss() # optimizing the images for e in range(args.iters): utils.imagenet_clamp_batch(output.data(), 0, 255) # fix BN for pre-trained vgg with autograd.record(): features_y = vgg(output.data()) content_loss = 2 * args.content_weight * mse_loss(features_y[1], f_xc_c) style_loss = 0. for m in range(len(features_y)): gram_y = net.gram_matrix(features_y[m]) gram_s = gram_style[m] style_loss = style_loss + 2 * args.style_weight * mse_loss(gram_y, gram_s) total_loss = content_loss + style_loss total_loss.backward() trainer.step(1) if (e + 1) % args.log_interval == 0: print('loss:{:.2f}'.format(total_loss.asnumpy()[0])) # save the image output = utils.add_imagenet_mean_batch(output.data()) utils.tensor_save_bgrimage(output[0], args.output_image, args.cuda)
Get symbol of mnist def get_mnist_sym(output_op=None, num_hidden=400): """Get symbol of mnist""" net = mx.symbol.Variable('data') net = mx.symbol.FullyConnected(data=net, name='mnist_fc1', num_hidden=num_hidden) net = mx.symbol.Activation(data=net, name='mnist_relu1', act_type="relu") net = mx.symbol.FullyConnected(data=net, name='mnist_fc2', num_hidden=num_hidden) net = mx.symbol.Activation(data=net, name='mnist_relu2', act_type="relu") net = mx.symbol.FullyConnected(data=net, name='mnist_fc3', num_hidden=10) if output_op is None: net = mx.symbol.SoftmaxOutput(data=net, name='softmax') else: net = output_op(data=net, name='softmax') return net
Get synthetic gradient value def synthetic_grad(X, theta, sigma1, sigma2, sigmax, rescale_grad=1.0, grad=None): """Get synthetic gradient value""" if grad is None: grad = nd.empty(theta.shape, theta.context) theta1 = theta.asnumpy()[0] theta2 = theta.asnumpy()[1] v1 = sigma1 ** 2 v2 = sigma2 ** 2 vx = sigmax ** 2 denominator = numpy.exp(-(X - theta1) ** 2 / (2 * vx)) + numpy.exp( -(X - theta1 - theta2) ** 2 / (2 * vx)) grad_npy = numpy.zeros(theta.shape) grad_npy[0] = -rescale_grad * ((numpy.exp(-(X - theta1) ** 2 / (2 * vx)) * (X - theta1) / vx + numpy.exp(-(X - theta1 - theta2) ** 2 / (2 * vx)) * (X - theta1 - theta2) / vx) / denominator).sum() + theta1 / v1 grad_npy[1] = -rescale_grad * ((numpy.exp(-(X - theta1 - theta2) ** 2 / (2 * vx)) * (X - theta1 - theta2) / vx) / denominator).sum() + theta2 / v2 grad[:] = grad_npy return grad
Get toy symbol def get_toy_sym(teacher=True, teacher_noise_precision=None): """Get toy symbol""" if teacher: net = mx.symbol.Variable('data') net = mx.symbol.FullyConnected(data=net, name='teacher_fc1', num_hidden=100) net = mx.symbol.Activation(data=net, name='teacher_relu1', act_type="relu") net = mx.symbol.FullyConnected(data=net, name='teacher_fc2', num_hidden=1) net = mx.symbol.LinearRegressionOutput(data=net, name='teacher_output', grad_scale=teacher_noise_precision) else: net = mx.symbol.Variable('data') net = mx.symbol.FullyConnected(data=net, name='student_fc1', num_hidden=100) net = mx.symbol.Activation(data=net, name='student_relu1', act_type="relu") student_mean = mx.symbol.FullyConnected(data=net, name='student_mean', num_hidden=1) student_var = mx.symbol.FullyConnected(data=net, name='student_var', num_hidden=1) net = mx.symbol.Group([student_mean, student_var]) return net
Run DistilledSGLD on mnist dataset def run_mnist_DistilledSGLD(num_training=50000, gpu_id=None): """Run DistilledSGLD on mnist dataset""" X, Y, X_test, Y_test = load_mnist(num_training) minibatch_size = 100 if num_training >= 10000: num_hidden = 800 total_iter_num = 1000000 teacher_learning_rate = 1E-6 student_learning_rate = 0.0001 teacher_prior = 1 student_prior = 0.1 perturb_deviation = 0.1 else: num_hidden = 400 total_iter_num = 20000 teacher_learning_rate = 4E-5 student_learning_rate = 0.0001 teacher_prior = 1 student_prior = 0.1 perturb_deviation = 0.001 teacher_net = get_mnist_sym(num_hidden=num_hidden) logsoftmax = LogSoftmax() student_net = get_mnist_sym(output_op=logsoftmax, num_hidden=num_hidden) data_shape = (minibatch_size,) + X.shape[1::] teacher_data_inputs = {'data': nd.zeros(data_shape, ctx=dev(gpu_id)), 'softmax_label': nd.zeros((minibatch_size,), ctx=dev(gpu_id))} student_data_inputs = {'data': nd.zeros(data_shape, ctx=dev(gpu_id)), 'softmax_label': nd.zeros((minibatch_size, 10), ctx=dev(gpu_id))} teacher_initializer = BiasXavier(factor_type="in", magnitude=1) student_initializer = BiasXavier(factor_type="in", magnitude=1) student_exe, student_params, _ = \ DistilledSGLD(teacher_sym=teacher_net, student_sym=student_net, teacher_data_inputs=teacher_data_inputs, student_data_inputs=student_data_inputs, X=X, Y=Y, X_test=X_test, Y_test=Y_test, total_iter_num=total_iter_num, student_initializer=student_initializer, teacher_initializer=teacher_initializer, student_optimizing_algorithm="adam", teacher_learning_rate=teacher_learning_rate, student_learning_rate=student_learning_rate, teacher_prior_precision=teacher_prior, student_prior_precision=student_prior, perturb_deviation=perturb_deviation, minibatch_size=100, dev=dev(gpu_id))
Run SGLD on toy dataset def run_toy_SGLD(gpu_id=None): """Run SGLD on toy dataset""" X, Y, X_test, Y_test = load_toy() minibatch_size = 1 teacher_noise_precision = 1.0 / 9.0 net = get_toy_sym(True, teacher_noise_precision) data_shape = (minibatch_size,) + X.shape[1::] data_inputs = {'data': nd.zeros(data_shape, ctx=dev(gpu_id)), 'teacher_output_label': nd.zeros((minibatch_size, 1), ctx=dev(gpu_id))} initializer = mx.init.Uniform(0.07) exe, params, _ = SGLD(sym=net, data_inputs=data_inputs, X=X, Y=Y, X_test=X_test, Y_test=Y_test, total_iter_num=50000, initializer=initializer, learning_rate=1E-4, # lr_scheduler=mx.lr_scheduler.FactorScheduler(100000, 0.5), prior_precision=0.1, burn_in_iter_num=1000, thin_interval=10, task='regression', minibatch_size=minibatch_size, dev=dev(gpu_id))
Run DistilledSGLD on toy dataset def run_toy_DistilledSGLD(gpu_id): """Run DistilledSGLD on toy dataset""" X, Y, X_test, Y_test = load_toy() minibatch_size = 1 teacher_noise_precision = 1.0 teacher_net = get_toy_sym(True, teacher_noise_precision) student_net = get_toy_sym(False) data_shape = (minibatch_size,) + X.shape[1::] teacher_data_inputs = {'data': nd.zeros(data_shape, ctx=dev(gpu_id)), 'teacher_output_label': nd.zeros((minibatch_size, 1), ctx=dev(gpu_id))} student_data_inputs = {'data': nd.zeros(data_shape, ctx=dev(gpu_id))} teacher_initializer = mx.init.Uniform(0.07) student_initializer = mx.init.Uniform(0.07) student_grad_f = lambda student_outputs, teacher_pred: \ regression_student_grad(student_outputs, teacher_pred, teacher_noise_precision) student_exe, student_params, _ = \ DistilledSGLD(teacher_sym=teacher_net, student_sym=student_net, teacher_data_inputs=teacher_data_inputs, student_data_inputs=student_data_inputs, X=X, Y=Y, X_test=X_test, Y_test=Y_test, total_iter_num=80000, teacher_initializer=teacher_initializer, student_initializer=student_initializer, teacher_learning_rate=1E-4, student_learning_rate=0.01, # teacher_lr_scheduler=mx.lr_scheduler.FactorScheduler(100000, 0.5), student_lr_scheduler=mx.lr_scheduler.FactorScheduler(8000, 0.8), student_grad_f=student_grad_f, teacher_prior_precision=0.1, student_prior_precision=0.001, perturb_deviation=0.1, minibatch_size=minibatch_size, task='regression', dev=dev(gpu_id))
Run HMC on toy dataset def run_toy_HMC(gpu_id=None): """Run HMC on toy dataset""" X, Y, X_test, Y_test = load_toy() minibatch_size = Y.shape[0] noise_precision = 1 / 9.0 net = get_toy_sym(True, noise_precision) data_shape = (minibatch_size,) + X.shape[1::] data_inputs = {'data': nd.zeros(data_shape, ctx=dev(gpu_id)), 'teacher_output_label': nd.zeros((minibatch_size, 1), ctx=dev(gpu_id))} initializer = mx.init.Uniform(0.07) sample_pool = HMC(net, data_inputs=data_inputs, X=X, Y=Y, X_test=X_test, Y_test=Y_test, sample_num=300000, initializer=initializer, prior_precision=1.0, learning_rate=1E-3, L=10, dev=dev(gpu_id))
Run synthetic SGLD def run_synthetic_SGLD(): """Run synthetic SGLD""" theta1 = 0 theta2 = 1 sigma1 = numpy.sqrt(10) sigma2 = 1 sigmax = numpy.sqrt(2) X = load_synthetic(theta1=theta1, theta2=theta2, sigmax=sigmax, num=100) minibatch_size = 1 total_iter_num = 1000000 lr_scheduler = SGLDScheduler(begin_rate=0.01, end_rate=0.0001, total_iter_num=total_iter_num, factor=0.55) optimizer = mx.optimizer.create('sgld', learning_rate=None, rescale_grad=1.0, lr_scheduler=lr_scheduler, wd=0) updater = mx.optimizer.get_updater(optimizer) theta = mx.random.normal(0, 1, (2,), mx.cpu()) grad = nd.empty((2,), mx.cpu()) samples = numpy.zeros((2, total_iter_num)) start = time.time() for i in range(total_iter_num): if (i + 1) % 100000 == 0: end = time.time() print("Iter:%d, Time spent: %f" % (i + 1, end - start)) start = time.time() ind = numpy.random.randint(0, X.shape[0]) synthetic_grad(X[ind], theta, sigma1, sigma2, sigmax, rescale_grad=X.shape[0] / float(minibatch_size), grad=grad) updater('theta', grad, theta) samples[:, i] = theta.asnumpy() plt.hist2d(samples[0, :], samples[1, :], (200, 200), cmap=plt.cm.jet) plt.colorbar() plt.show()
wrapper function for loading pascal voc dataset Parameters: ---------- image_set : str train, trainval... year : str 2007, 2012 or combinations splitted by comma devkit_path : str root directory of dataset shuffle : bool whether to shuffle initial list Returns: ---------- Imdb def load_pascal(image_set, year, devkit_path, shuffle=False): """ wrapper function for loading pascal voc dataset Parameters: ---------- image_set : str train, trainval... year : str 2007, 2012 or combinations splitted by comma devkit_path : str root directory of dataset shuffle : bool whether to shuffle initial list Returns: ---------- Imdb """ image_set = [y.strip() for y in image_set.split(',')] assert image_set, "No image_set specified" year = [y.strip() for y in year.split(',')] assert year, "No year specified" # make sure (# sets == # years) if len(image_set) > 1 and len(year) == 1: year = year * len(image_set) if len(image_set) == 1 and len(year) > 1: image_set = image_set * len(year) assert len(image_set) == len(year), "Number of sets and year mismatch" imdbs = [] for s, y in zip(image_set, year): imdbs.append(PascalVoc(s, y, devkit_path, shuffle, is_train=True)) if len(imdbs) > 1: return ConcatDB(imdbs, shuffle) else: return imdbs[0]
wrapper function for loading ms coco dataset Parameters: ---------- image_set : str train2014, val2014, valminusminival2014, minival2014 dirname: str root dir for coco shuffle: boolean initial shuffle def load_coco(image_set, dirname, shuffle=False): """ wrapper function for loading ms coco dataset Parameters: ---------- image_set : str train2014, val2014, valminusminival2014, minival2014 dirname: str root dir for coco shuffle: boolean initial shuffle """ anno_files = ['instances_' + y.strip() + '.json' for y in image_set.split(',')] assert anno_files, "No image set specified" imdbs = [] for af in anno_files: af_path = os.path.join(dirname, 'annotations', af) imdbs.append(Coco(af_path, dirname, shuffle=shuffle)) if len(imdbs) > 1: return ConcatDB(imdbs, shuffle) else: return imdbs[0]
Resets the iterator to the beginning of the data. def reset(self): """Resets the iterator to the beginning of the data.""" self.curr_idx = 0 #shuffle data in each bucket random.shuffle(self.idx) for i, buck in enumerate(self.sentences): self.indices[i], self.sentences[i], self.characters[i], self.label[i] = shuffle(self.indices[i], self.sentences[i], self.characters[i], self.label[i]) self.ndindex = [] self.ndsent = [] self.ndchar = [] self.ndlabel = [] #for each bucket of data for i, buck in enumerate(self.sentences): #append the lists with an array self.ndindex.append(ndarray.array(self.indices[i], dtype=self.dtype)) self.ndsent.append(ndarray.array(self.sentences[i], dtype=self.dtype)) self.ndchar.append(ndarray.array(self.characters[i], dtype=self.dtype)) self.ndlabel.append(ndarray.array(self.label[i], dtype=self.dtype))
Returns the next batch of data. def next(self): """Returns the next batch of data.""" if self.curr_idx == len(self.idx): raise StopIteration #i = batches index, j = starting record i, j = self.idx[self.curr_idx] self.curr_idx += 1 indices = self.ndindex[i][j:j + self.batch_size] sentences = self.ndsent[i][j:j + self.batch_size] characters = self.ndchar[i][j:j + self.batch_size] label = self.ndlabel[i][j:j + self.batch_size] return DataBatch([sentences, characters], [label], pad=0, index = indices, bucket_key=self.buckets[i], provide_data=[DataDesc(name=self.data_names[0], shape=sentences.shape, layout=self.layout), DataDesc(name=self.data_names[1], shape=characters.shape, layout=self.layout)], provide_label=[DataDesc(name=self.label_name, shape=label.shape, layout=self.layout)])
Converts a reshape layer from mxnet to coreml. This doesn't currently handle the deprecated parameters for the reshape layer. Parameters ---------- network: net An mxnet network object. layer: node Node to convert. module: module A module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_reshape(net, node, module, builder): """Converts a reshape layer from mxnet to coreml. This doesn't currently handle the deprecated parameters for the reshape layer. Parameters ---------- network: net An mxnet network object. layer: node Node to convert. module: module A module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] target_shape = node['shape'] if any(item <= 0 for item in target_shape): raise NotImplementedError('Special dimensional values less than or equal to 0 are not supported yet.' 'Feel free to file an issue here: https://github.com/dmlc/mxnet/issues.') if 'reverse' in node and node['reverse'] == 'True': raise NotImplementedError('"reverse" parameter is not supported by yet.' 'Feel free to file an issue here: https://github.com/dmlc/mxnet/issues.') mode = 0 # CHANNEL_FIRST builder.add_reshape(name, input_name, output_name, target_shape, mode)
Convert a transpose layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_transpose(net, node, module, builder): """Convert a transpose layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] param = _get_attrs(node) axes = literal_eval(param['axes']) builder.add_permute(name, axes, input_name, output_name)
Convert a flatten layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_flatten(net, node, module, builder): """Convert a flatten layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] mode = 0 # CHANNEL_FIRST builder.add_flatten(name, mode, input_name, output_name)
Convert a softmax layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_softmax(net, node, module, builder): """Convert a softmax layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] builder.add_softmax(name=name, input_name=input_name, output_name=output_name)
Convert an activation layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_activation(net, node, module, builder): """Convert an activation layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] mx_non_linearity = _get_attrs(node)['act_type'] #TODO add SCALED_TANH, SOFTPLUS, SOFTSIGN, SIGMOID_HARD, LEAKYRELU, PRELU, ELU, PARAMETRICSOFTPLUS, THRESHOLDEDRELU, LINEAR if mx_non_linearity == 'relu': non_linearity = 'RELU' elif mx_non_linearity == 'tanh': non_linearity = 'TANH' elif mx_non_linearity == 'sigmoid': non_linearity = 'SIGMOID' else: raise TypeError('Unknown activation type %s' % mx_non_linearity) builder.add_activation(name = name, non_linearity = non_linearity, input_name = input_name, output_name = output_name)
Convert a leakyrelu layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_leakyrelu(net, node, module, builder): """Convert a leakyrelu layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] inputs = node['inputs'] args, _ = module.get_params() mx_non_linearity = _get_attrs(node)['act_type'] if mx_non_linearity == 'elu': non_linearity = 'ELU' slope = _get_attrs(node)['slope'] if 'slope' in _get_attrs(node) else 0.25 params = slope elif mx_non_linearity == 'leaky': non_linearity = 'LEAKYRELU' slope = _get_attrs(node)['slope'] if 'slope' in _get_attrs(node) else 0.25 params = [slope] elif mx_non_linearity == 'prelu': non_linearity = 'PRELU' params = args[_get_node_name(net, inputs[1][0])].asnumpy() else: raise TypeError('Unknown activation type %s' % mx_non_linearity) builder.add_activation(name = name, non_linearity = non_linearity, input_name = input_name, output_name = output_name, params = params)
Convert an elementwise add layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_elementwise_add(net, node, module, builder): """Convert an elementwise add layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_names, output_name = _get_input_output_name(net, node, [0, 1]) name = node['name'] builder.add_elementwise(name, input_names, output_name, 'ADD')
Convert a convolution layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_convolution(net, node, module, builder): """Convert a convolution layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] param = _get_attrs(node) inputs = node['inputs'] args, _ = module.get_params() if 'no_bias' in param.keys(): has_bias = not literal_eval(param['no_bias']) else: has_bias = True if 'pad' in param.keys() and literal_eval(param['pad']) != (0, 0): pad = literal_eval(param['pad']) builder.add_padding( name=name+"_pad", left=pad[1], right=pad[1], top=pad[0], bottom=pad[0], value=0, input_name=input_name, output_name=name+"_pad_output") input_name = name+"_pad_output" border_mode = "valid" n_filters = int(param['num_filter']) n_groups = int(param['num_group']) if 'num_group' in param else 1 W = args[_get_node_name(net, inputs[1][0])].asnumpy() if has_bias: Wb = args[_get_node_name(net, inputs[2][0])].asnumpy() else: Wb = None channels = W.shape[1] stride_height = 1 stride_width = 1 if 'stride' in param.keys(): stride_height, stride_width = literal_eval(param['stride']) kernel_height, kernel_width = literal_eval(param['kernel']) W = W.transpose((2, 3, 1, 0)) builder.add_convolution( name=name, kernel_channels=channels, output_channels=n_filters, height=kernel_height, width=kernel_width, stride_height=stride_height, stride_width=stride_width, border_mode=border_mode, groups=n_groups, W=W, b=Wb, has_bias=has_bias, is_deconv=False, output_shape=None, input_name=input_name, output_name=output_name)
Convert a pooling layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_pooling(net, node, module, builder): """Convert a pooling layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] param = _get_attrs(node) layer_type_mx = param['pool_type'] if layer_type_mx == 'max': layer_type = 'MAX' elif layer_type_mx == 'avg': layer_type = 'AVERAGE' else: raise TypeError("Pooling type %s not supported" % layer_type_mx) # Add padding if there is any if 'pad' in param.keys() and literal_eval(param['pad']) != (0, 0): pad = literal_eval(param['pad']) builder.add_padding( name=name+"_pad", left=pad[1], right=pad[1], top=pad[0], bottom=pad[0], value=0, input_name=input_name, output_name=name+"_pad_output") input_name = name+"_pad_output" stride_height = 1 stride_width = 1 if 'stride' in param.keys(): stride_height, stride_width = literal_eval(param['stride']) kernel_width, kernel_height = literal_eval(param['kernel']) type_map = {'valid': 'VALID', 'full': 'INCLUDE_LAST_PIXEL'} padding_type = param['pooling_convention'] if 'pooling_convention' in param else 'valid' if padding_type not in type_map: raise KeyError("%s type is not supported in this converter. It is a Github issue.") padding_type = type_map[padding_type] if 'global_pool' in param.keys(): is_global = literal_eval(param['global_pool']) else: is_global = False # For reasons why we are not using the standard builder but having our own implementation, # see the function documentation. _add_pooling.add_pooling_with_padding_types( builder=builder, name=name, height=kernel_height, width=kernel_width, stride_height=stride_height, stride_width=stride_width, layer_type=layer_type, padding_type=padding_type, exclude_pad_area=False, is_global=is_global, input_name=input_name, output_name=output_name )
Convert a batchnorm layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_batchnorm(net, node, module, builder): """Convert a batchnorm layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ input_name, output_name = _get_input_output_name(net, node) name = node['name'] inputs = node['inputs'] eps = 1e-3 # Default value of eps for MXNet. use_global_stats = False # Default value of use_global_stats for MXNet. fix_gamma = True # Default value of fix_gamma for MXNet. attrs = _get_attrs(node) if 'eps' in attrs: eps = literal_eval(attrs['eps']) if 'fix_gamma' in attrs: fix_gamma = literal_eval(attrs['fix_gamma']) args, aux = module.get_params() gamma = args[_get_node_name(net, inputs[1][0])].asnumpy() beta = args[_get_node_name(net, inputs[2][0])].asnumpy() mean = aux[_get_node_name(net, inputs[3][0])].asnumpy() variance = aux[_get_node_name(net, inputs[4][0])].asnumpy() nb_channels = gamma.shape[0] if fix_gamma: gamma.fill(1.) builder.add_batchnorm( name=name, channels=nb_channels, gamma=gamma, beta=beta, mean=mean, variance=variance, input_name=input_name, output_name=output_name, epsilon=eps)
Convert concat layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. def convert_concat(net, node, module, builder): """Convert concat layer from mxnet to coreml. Parameters ---------- network: net A mxnet network object. layer: node Node to convert. module: module An module for MXNet builder: NeuralNetworkBuilder A neural network builder object. """ # Get input and output names input_names, output_name = _get_input_output_name(net, node, 'all') name = node['name'] mode = 'CONCAT' builder.add_elementwise(name = name, input_names = input_names, output_name = output_name, mode = mode)
convert from mxnet's opts to dmlc's opts def dmlc_opts(opts): """convert from mxnet's opts to dmlc's opts """ args = ['--num-workers', str(opts.num_workers), '--num-servers', str(opts.num_servers), '--cluster', opts.launcher, '--host-file', opts.hostfile, '--sync-dst-dir', opts.sync_dst_dir] # convert to dictionary dopts = vars(opts) for key in ['env_server', 'env_worker', 'env']: for v in dopts[key]: args.append('--' + key.replace("_","-")) args.append(v) args += opts.command try: from dmlc_tracker import opts except ImportError: print("Can't load dmlc_tracker package. Perhaps you need to run") print(" git submodule update --init --recursive") raise dmlc_opts = opts.get_opts(args) return dmlc_opts
Unfuses the fused RNN in to a stack of rnn cells. def _unfuse(self): """Unfuses the fused RNN in to a stack of rnn cells.""" assert not self._projection_size, "_unfuse does not support projection layer yet!" assert not self._lstm_state_clip_min and not self._lstm_state_clip_max, \ "_unfuse does not support state clipping yet!" get_cell = {'rnn_relu': lambda **kwargs: rnn_cell.RNNCell(self._hidden_size, activation='relu', **kwargs), 'rnn_tanh': lambda **kwargs: rnn_cell.RNNCell(self._hidden_size, activation='tanh', **kwargs), 'lstm': lambda **kwargs: rnn_cell.LSTMCell(self._hidden_size, **kwargs), 'gru': lambda **kwargs: rnn_cell.GRUCell(self._hidden_size, **kwargs)}[self._mode] stack = rnn_cell.HybridSequentialRNNCell(prefix=self.prefix, params=self.params) with stack.name_scope(): ni = self._input_size for i in range(self._num_layers): kwargs = {'input_size': ni, 'i2h_weight_initializer': self._i2h_weight_initializer, 'h2h_weight_initializer': self._h2h_weight_initializer, 'i2h_bias_initializer': self._i2h_bias_initializer, 'h2h_bias_initializer': self._h2h_bias_initializer} if self._dir == 2: stack.add(rnn_cell.BidirectionalCell( get_cell(prefix='l%d_'%i, **kwargs), get_cell(prefix='r%d_'%i, **kwargs))) else: stack.add(get_cell(prefix='l%d_'%i, **kwargs)) if self._dropout > 0 and i != self._num_layers - 1: stack.add(rnn_cell.DropoutCell(self._dropout)) ni = self._hidden_size * self._dir return stack
Initial state for this cell. Parameters ---------- batch_size: int Only required for `NDArray` API. Size of the batch ('N' in layout). Dimension of the input. func : callable, default `ndarray.zeros` Function for creating initial state. For Symbol API, func can be `symbol.zeros`, `symbol.uniform`, `symbol.var` etc. Use `symbol.var` if you want to directly feed input as states. For NDArray API, func can be `ndarray.zeros`, `ndarray.ones`, etc. **kwargs : Additional keyword arguments passed to func. For example `mean`, `std`, `dtype`, etc. Returns ------- states : nested list of Symbol Starting states for the first RNN step. def begin_state(self, batch_size=0, func=ndarray.zeros, **kwargs): """Initial state for this cell. Parameters ---------- batch_size: int Only required for `NDArray` API. Size of the batch ('N' in layout). Dimension of the input. func : callable, default `ndarray.zeros` Function for creating initial state. For Symbol API, func can be `symbol.zeros`, `symbol.uniform`, `symbol.var` etc. Use `symbol.var` if you want to directly feed input as states. For NDArray API, func can be `ndarray.zeros`, `ndarray.ones`, etc. **kwargs : Additional keyword arguments passed to func. For example `mean`, `std`, `dtype`, etc. Returns ------- states : nested list of Symbol Starting states for the first RNN step. """ states = [] for i, info in enumerate(self.state_info(batch_size)): if info is not None: info.update(kwargs) else: info = kwargs states.append(func(name='%sh0_%d'%(self.prefix, i), **info)) return states
forward using CUDNN or CPU kenrel def _forward_kernel(self, F, inputs, states, **kwargs): """ forward using CUDNN or CPU kenrel""" if self._layout == 'NTC': inputs = F.swapaxes(inputs, dim1=0, dim2=1) if self._projection_size is None: params = (kwargs['{}{}_{}_{}'.format(d, l, g, t)].reshape(-1) for t in ['weight', 'bias'] for l in range(self._num_layers) for d in ['l', 'r'][:self._dir] for g in ['i2h', 'h2h']) else: params = (kwargs['{}{}_{}_{}'.format(d, l, g, t)].reshape(-1) for t in ['weight', 'bias'] for l in range(self._num_layers) for d in ['l', 'r'][:self._dir] for g in ['i2h', 'h2h', 'h2r'] if g != 'h2r' or t != 'bias') params = F._internal._rnn_param_concat(*params, dim=0) rnn = F.RNN(inputs, params, *states, state_size=self._hidden_size, projection_size=self._projection_size, num_layers=self._num_layers, bidirectional=self._dir == 2, p=self._dropout, state_outputs=True, mode=self._mode, lstm_state_clip_min=self._lstm_state_clip_min, lstm_state_clip_max=self._lstm_state_clip_max, lstm_state_clip_nan=self._lstm_state_clip_nan) if self._mode == 'lstm': outputs, states = rnn[0], [rnn[1], rnn[2]] else: outputs, states = rnn[0], [rnn[1]] if self._layout == 'NTC': outputs = F.swapaxes(outputs, dim1=0, dim2=1) return outputs, states
Wait for network service to appear @param server: host to connect to (str) @param port: port (int) @param timeout: in seconds, if None or 0 wait forever @return: True of False, if timeout is None may return only True or throw unhandled network exception def wait_ssh_open(server, port, keep_waiting=None, timeout=None): """ Wait for network service to appear @param server: host to connect to (str) @param port: port (int) @param timeout: in seconds, if None or 0 wait forever @return: True of False, if timeout is None may return only True or throw unhandled network exception """ import socket import errno import time log = logging.getLogger('wait_ssh_open') sleep_s = 1 if timeout: from time import time as now # time module is needed to calc timeout shared between two exceptions end = now() + timeout while True: log.debug("Sleeping for %s second(s)", sleep_s) time.sleep(sleep_s) s = socket.socket() try: if keep_waiting and not keep_waiting(): log.debug("keep_waiting() is set and evaluates to False") return False if timeout: next_timeout = end - now() if next_timeout < 0: log.debug("connect time out") return False else: log.debug("connect timeout %d s", next_timeout) s.settimeout(next_timeout) log.debug("connect %s:%d", server, port) s.connect((server, port)) ret = s.recv(1024).decode() if ret and ret.startswith('SSH'): s.close() log.info("wait_ssh_open: port %s:%s is open and ssh is ready", server, port) return True else: log.debug("Didn't get the SSH banner") s.close() except ConnectionError as err: log.debug("ConnectionError %s", err) if sleep_s == 0: sleep_s = 1 else: sleep_s *= 2 except socket.gaierror as err: log.debug("gaierror %s",err) return False except socket.timeout as err: # this exception occurs only if timeout is set if timeout: return False except TimeoutError as err: # catch timeout exception from underlying network library # this one is different from socket.timeout raise
Wait for network service to appear @param server: host to connect to (str) @param port: port (int) @param timeout: in seconds, if None or 0 wait forever @return: True of False, if timeout is None may return only True or throw unhandled network exception def wait_port_open(server, port, timeout=None): """ Wait for network service to appear @param server: host to connect to (str) @param port: port (int) @param timeout: in seconds, if None or 0 wait forever @return: True of False, if timeout is None may return only True or throw unhandled network exception """ import socket import errno import time sleep_s = 0 if timeout: from time import time as now # time module is needed to calc timeout shared between two exceptions end = now() + timeout while True: logging.debug("Sleeping for %s second(s)", sleep_s) time.sleep(sleep_s) s = socket.socket() try: if timeout: next_timeout = end - now() if next_timeout < 0: return False else: s.settimeout(next_timeout) logging.info("connect %s %d", server, port) s.connect((server, port)) except ConnectionError as err: logging.debug("ConnectionError %s", err) if sleep_s == 0: sleep_s = 1 except socket.gaierror as err: logging.debug("gaierror %s",err) return False except socket.timeout as err: # this exception occurs only if timeout is set if timeout: return False except TimeoutError as err: # catch timeout exception from underlying network library # this one is different from socket.timeout raise else: s.close() logging.info("wait_port_open: port %s:%s is open", server, port) return True
Convert symbol for detail information. Parameters ---------- symbol: Symbol Symbol to be visualized. shape: dict A dict of shapes, str->shape (tuple), given input shapes. line_length: int Rotal length of printed lines positions: list Relative or absolute positions of log elements in each line. Returns ------ None Notes ----- If ``mxnet`` is imported, the visualization module can be used in its short-form. For example, if we ``import mxnet`` as follows:: import mxnet this method in visualization module can be used in its short-form as:: mxnet.viz.print_summary(...) def print_summary(symbol, shape=None, line_length=120, positions=[.44, .64, .74, 1.]): """Convert symbol for detail information. Parameters ---------- symbol: Symbol Symbol to be visualized. shape: dict A dict of shapes, str->shape (tuple), given input shapes. line_length: int Rotal length of printed lines positions: list Relative or absolute positions of log elements in each line. Returns ------ None Notes ----- If ``mxnet`` is imported, the visualization module can be used in its short-form. For example, if we ``import mxnet`` as follows:: import mxnet this method in visualization module can be used in its short-form as:: mxnet.viz.print_summary(...) """ if not isinstance(symbol, Symbol): raise TypeError("symbol must be Symbol") show_shape = False if shape is not None: show_shape = True interals = symbol.get_internals() _, out_shapes, _ = interals.infer_shape(**shape) if out_shapes is None: raise ValueError("Input shape is incomplete") shape_dict = dict(zip(interals.list_outputs(), out_shapes)) conf = json.loads(symbol.tojson()) nodes = conf["nodes"] heads = set(conf["heads"][0]) if positions[-1] <= 1: positions = [int(line_length * p) for p in positions] # header names for the different log elements to_display = ['Layer (type)', 'Output Shape', 'Param #', 'Previous Layer'] def print_row(fields, positions): """Print format row. Parameters ---------- fields: list Information field. positions: list Field length ratio. Returns ------ None """ line = '' for i, field in enumerate(fields): line += str(field) line = line[:positions[i]] line += ' ' * (positions[i] - len(line)) print(line) print('_' * line_length) print_row(to_display, positions) print('=' * line_length) def print_layer_summary(node, out_shape): """print layer information Parameters ---------- node: dict Node information. out_shape: dict Node shape information. Returns ------ Node total parameters. """ op = node["op"] pre_node = [] pre_filter = 0 if op != "null": inputs = node["inputs"] for item in inputs: input_node = nodes[item[0]] input_name = input_node["name"] if input_node["op"] != "null" or item[0] in heads: # add precede pre_node.append(input_name) if show_shape: if input_node["op"] != "null": key = input_name + "_output" else: key = input_name if key in shape_dict: shape = shape_dict[key][1:] pre_filter = pre_filter + int(shape[0]) cur_param = 0 if op == 'Convolution': if "no_bias" in node["attrs"] and node["attrs"]["no_bias"] == 'True': num_group = int(node['attrs'].get('num_group', '1')) cur_param = pre_filter * int(node["attrs"]["num_filter"]) \ // num_group for k in _str2tuple(node["attrs"]["kernel"]): cur_param *= int(k) else: num_group = int(node['attrs'].get('num_group', '1')) cur_param = pre_filter * int(node["attrs"]["num_filter"]) \ // num_group for k in _str2tuple(node["attrs"]["kernel"]): cur_param *= int(k) cur_param += int(node["attrs"]["num_filter"]) elif op == 'FullyConnected': if "no_bias" in node["attrs"] and node["attrs"]["no_bias"] == 'True': cur_param = pre_filter * int(node["attrs"]["num_hidden"]) else: cur_param = (pre_filter+1) * int(node["attrs"]["num_hidden"]) elif op == 'BatchNorm': key = node["name"] + "_output" if show_shape: num_filter = shape_dict[key][1] cur_param = int(num_filter) * 2 elif op == 'Embedding': cur_param = int(node["attrs"]['input_dim']) * int(node["attrs"]['output_dim']) if not pre_node: first_connection = '' else: first_connection = pre_node[0] fields = [node['name'] + '(' + op + ')', "x".join([str(x) for x in out_shape]), cur_param, first_connection] print_row(fields, positions) if len(pre_node) > 1: for i in range(1, len(pre_node)): fields = ['', '', '', pre_node[i]] print_row(fields, positions) return cur_param total_params = 0 for i, node in enumerate(nodes): out_shape = [] op = node["op"] if op == "null" and i > 0: continue if op != "null" or i in heads: if show_shape: if op != "null": key = node["name"] + "_output" else: key = node["name"] if key in shape_dict: out_shape = shape_dict[key][1:] total_params += print_layer_summary(nodes[i], out_shape) if i == len(nodes) - 1: print('=' * line_length) else: print('_' * line_length) print("Total params: {params}".format(params=total_params)) print('_' * line_length)
Creates a visualization (Graphviz digraph object) of the given computation graph. Graphviz must be installed for this function to work. Parameters ---------- title: str, optional Title of the generated visualization. symbol: Symbol A symbol from the computation graph. The generated digraph will visualize the part of the computation graph required to compute `symbol`. shape: dict, optional Specifies the shape of the input tensors. If specified, the visualization will include the shape of the tensors between the nodes. `shape` is a dictionary mapping input symbol names (str) to the corresponding tensor shape (tuple). dtype: dict, optional Specifies the type of the input tensors. If specified, the visualization will include the type of the tensors between the nodes. `dtype` is a dictionary mapping input symbol names (str) to the corresponding tensor type (e.g. `numpy.float32`). node_attrs: dict, optional Specifies the attributes for nodes in the generated visualization. `node_attrs` is a dictionary of Graphviz attribute names and values. For example:: node_attrs={"shape":"oval","fixedsize":"false"} will use oval shape for nodes and allow variable sized nodes in the visualization. hide_weights: bool, optional If True (default), then inputs with names of form *_weight* (corresponding to weight tensors) or *_bias* (corresponding to bias vectors) will be hidden for a cleaner visualization. Returns ------- dot: Digraph A Graphviz digraph object visualizing the computation graph to compute `symbol`. Example ------- >>> net = mx.sym.Variable('data') >>> net = mx.sym.FullyConnected(data=net, name='fc1', num_hidden=128) >>> net = mx.sym.Activation(data=net, name='relu1', act_type="relu") >>> net = mx.sym.FullyConnected(data=net, name='fc2', num_hidden=10) >>> net = mx.sym.SoftmaxOutput(data=net, name='out') >>> digraph = mx.viz.plot_network(net, shape={'data':(100,200)}, ... node_attrs={"fixedsize":"false"}) >>> digraph.view() Notes ----- If ``mxnet`` is imported, the visualization module can be used in its short-form. For example, if we ``import mxnet`` as follows:: import mxnet this method in visualization module can be used in its short-form as:: mxnet.viz.plot_network(...) def plot_network(symbol, title="plot", save_format='pdf', shape=None, dtype=None, node_attrs={}, hide_weights=True): """Creates a visualization (Graphviz digraph object) of the given computation graph. Graphviz must be installed for this function to work. Parameters ---------- title: str, optional Title of the generated visualization. symbol: Symbol A symbol from the computation graph. The generated digraph will visualize the part of the computation graph required to compute `symbol`. shape: dict, optional Specifies the shape of the input tensors. If specified, the visualization will include the shape of the tensors between the nodes. `shape` is a dictionary mapping input symbol names (str) to the corresponding tensor shape (tuple). dtype: dict, optional Specifies the type of the input tensors. If specified, the visualization will include the type of the tensors between the nodes. `dtype` is a dictionary mapping input symbol names (str) to the corresponding tensor type (e.g. `numpy.float32`). node_attrs: dict, optional Specifies the attributes for nodes in the generated visualization. `node_attrs` is a dictionary of Graphviz attribute names and values. For example:: node_attrs={"shape":"oval","fixedsize":"false"} will use oval shape for nodes and allow variable sized nodes in the visualization. hide_weights: bool, optional If True (default), then inputs with names of form *_weight* (corresponding to weight tensors) or *_bias* (corresponding to bias vectors) will be hidden for a cleaner visualization. Returns ------- dot: Digraph A Graphviz digraph object visualizing the computation graph to compute `symbol`. Example ------- >>> net = mx.sym.Variable('data') >>> net = mx.sym.FullyConnected(data=net, name='fc1', num_hidden=128) >>> net = mx.sym.Activation(data=net, name='relu1', act_type="relu") >>> net = mx.sym.FullyConnected(data=net, name='fc2', num_hidden=10) >>> net = mx.sym.SoftmaxOutput(data=net, name='out') >>> digraph = mx.viz.plot_network(net, shape={'data':(100,200)}, ... node_attrs={"fixedsize":"false"}) >>> digraph.view() Notes ----- If ``mxnet`` is imported, the visualization module can be used in its short-form. For example, if we ``import mxnet`` as follows:: import mxnet this method in visualization module can be used in its short-form as:: mxnet.viz.plot_network(...) """ # todo add shape support try: from graphviz import Digraph except: raise ImportError("Draw network requires graphviz library") if not isinstance(symbol, Symbol): raise TypeError("symbol must be a Symbol") internals = symbol.get_internals() draw_shape = shape is not None if draw_shape: _, out_shapes, _ = internals.infer_shape(**shape) if out_shapes is None: raise ValueError("Input shape is incomplete") shape_dict = dict(zip(internals.list_outputs(), out_shapes)) draw_type = dtype is not None if draw_type: _, out_types, _ = internals.infer_type(**dtype) if out_types is None: raise ValueError("Input type is incomplete") type_dict = dict(zip(internals.list_outputs(), out_types)) conf = json.loads(symbol.tojson()) nodes = conf["nodes"] # check if multiple nodes have the same name if len(nodes) != len(set([node["name"] for node in nodes])): seen_nodes = set() # find all repeated names repeated = set(node['name'] for node in nodes if node['name'] in seen_nodes or seen_nodes.add(node['name'])) warning_message = "There are multiple variables with the same name in your graph, " \ "this may result in cyclic graph. Repeated names: " + ','.join(repeated) warnings.warn(warning_message, RuntimeWarning) # default attributes of node node_attr = {"shape": "box", "fixedsize": "true", "width": "1.3", "height": "0.8034", "style": "filled"} # merge the dict provided by user and the default one node_attr.update(node_attrs) dot = Digraph(name=title, format=save_format) # color map cm = ("#8dd3c7", "#fb8072", "#ffffb3", "#bebada", "#80b1d3", "#fdb462", "#b3de69", "#fccde5") def looks_like_weight(name): """Internal helper to figure out if node should be hidden with `hide_weights`. """ weight_like = ('_weight', '_bias', '_beta', '_gamma', '_moving_var', '_moving_mean', '_running_var', '_running_mean') return name.endswith(weight_like) # make nodes hidden_nodes = set() for node in nodes: op = node["op"] name = node["name"] # input data attr = copy.deepcopy(node_attr) label = name if op == "null": if looks_like_weight(node["name"]): if hide_weights: hidden_nodes.add(node["name"]) # else we don't render a node, but # don't add it to the hidden_nodes set # so it gets rendered as an empty oval continue attr["shape"] = "oval" # inputs get their own shape label = node["name"] attr["fillcolor"] = cm[0] elif op == "Convolution": label = "Convolution\n{kernel}/{stride}, {filter}".format( kernel="x".join(_str2tuple(node["attrs"]["kernel"])), stride="x".join(_str2tuple(node["attrs"]["stride"])) if "stride" in node["attrs"] else "1", filter=node["attrs"]["num_filter"] ) attr["fillcolor"] = cm[1] elif op == "FullyConnected": label = "FullyConnected\n{hidden}".format(hidden=node["attrs"]["num_hidden"]) attr["fillcolor"] = cm[1] elif op == "BatchNorm": attr["fillcolor"] = cm[3] elif op == 'Activation': act_type = node["attrs"]["act_type"] label = 'Activation\n{activation}'.format(activation=act_type) attr["fillcolor"] = cm[2] elif op == 'LeakyReLU': attrs = node.get("attrs") act_type = attrs.get("act_type", "Leaky") if attrs else "Leaky" label = 'LeakyReLU\n{activation}'.format(activation=act_type) attr["fillcolor"] = cm[2] elif op == "Pooling": label = "Pooling\n{pooltype}, {kernel}/{stride}".format(pooltype=node["attrs"]["pool_type"], kernel="x".join(_str2tuple(node["attrs"]["kernel"])) if "kernel" in node["attrs"] else "[]", stride="x".join(_str2tuple(node["attrs"]["stride"])) if "stride" in node["attrs"] else "1") attr["fillcolor"] = cm[4] elif op in ("Concat", "Flatten", "Reshape"): attr["fillcolor"] = cm[5] elif op == "Softmax": attr["fillcolor"] = cm[6] else: attr["fillcolor"] = cm[7] if op == "Custom": label = node["attrs"]["op_type"] dot.node(name=name, label=label, **attr) # add edges for node in nodes: # pylint: disable=too-many-nested-blocks op = node["op"] name = node["name"] if op == "null": continue else: inputs = node["inputs"] for item in inputs: input_node = nodes[item[0]] input_name = input_node["name"] if input_name not in hidden_nodes: attr = {"dir": "back", 'arrowtail':'open', 'label': ''} # add shapes if draw_shape: if input_node["op"] != "null": key = input_name + "_output" if "attrs" in input_node: params = input_node["attrs"] if "num_outputs" in params: key += str(int(params["num_outputs"]) - 1) shape = shape_dict[key][1:] label = "x".join([str(x) for x in shape]) attr["label"] = label else: key = input_name shape = shape_dict[key][1:] label = "x".join([str(x) for x in shape]) attr["label"] = label if draw_type: if input_node["op"] != "null": key = input_name + "_output" if "attrs" in input_node: params = input_node["attrs"] if "num_outputs" in params: key += str(int(params["num_outputs"]) - 1) dtype = type_dict[key] attr["label"] += '(' + dtype.__name__ + ')' else: key = input_name dtype = type_dict[key] attr["label"] += '(' + dtype.__name__ + ')' dot.edge(tail_name=name, head_name=input_name, **attr) return dot
Measure the accuracy of ResNet Parameters ---------- data_iterator: Iter examples of dataset network: ResNet Returns ---------- tuple of array element def evaluate_accuracy(data_iterator, network): """ Measure the accuracy of ResNet Parameters ---------- data_iterator: Iter examples of dataset network: ResNet Returns ---------- tuple of array element """ acc = mx.metric.Accuracy() # Iterate through data and label for i, (data, label) in enumerate(data_iterator): # Get the data and label into the GPU data = data.as_in_context(ctx[0]) label = label.as_in_context(ctx[0]) # Get network's output which is a probability distribution # Apply argmax on the probability distribution to get network's classification. output = network(data) predictions = nd.argmax(output, axis=1) # Give network's prediction and the correct label to update the metric acc.update(preds=predictions, labels=label) # Return the accuracy return acc.get()[1]
Training with multiple GPUs Parameters ---------- batch_list: List list of dataset context: List a list of all GPUs to be used for training network: ResNet gluon_trainer: rain module of gluon def train_batch(batch_list, context, network, gluon_trainer): """ Training with multiple GPUs Parameters ---------- batch_list: List list of dataset context: List a list of all GPUs to be used for training network: ResNet gluon_trainer: rain module of gluon """ # Split and load data into multiple GPUs data = batch_list[0] data = gluon.utils.split_and_load(data, context) # Split and load label into multiple GPUs label = batch_list[1] label = gluon.utils.split_and_load(label, context) # Run the forward and backward pass forward_backward(network, data, label) # Update the parameters this_batch_size = batch_list[0].shape[0] gluon_trainer.step(this_batch_size)
Take an executor's underlying symbol graph and return its generated optimized version. Parameters ---------- executor : An executor for which you want to see an optimized symbol. Getting an optimized symbol is useful to compare and verify the work TensorRT has done against a legacy behaviour. Returns ------- symbol : nnvm::Symbol The nnvm symbol optimized. def get_optimized_symbol(executor): """ Take an executor's underlying symbol graph and return its generated optimized version. Parameters ---------- executor : An executor for which you want to see an optimized symbol. Getting an optimized symbol is useful to compare and verify the work TensorRT has done against a legacy behaviour. Returns ------- symbol : nnvm::Symbol The nnvm symbol optimized. """ handle = SymbolHandle() try: check_call(_LIB.MXExecutorGetOptimizedSymbol(executor.handle, ctypes.byref(handle))) result = sym.Symbol(handle=handle) return result except MXNetError: logging.error('Error while trying to fetch TRT optimized symbol for graph. Please ensure ' 'build was compiled with MXNET_USE_TENSORRT enabled.') raise
Bind current symbol to get an optimized trt executor. Parameters ---------- symbol : Symbol The symbol you wish to bind, and optimize with TensorRT. ctx : Context The device context the generated executor to run on. all_params : Dict of str->ndarray A dictionary of mappings from parameter names to parameter NDArrays. type_dict : Dict of str->numpy.dtype Input type dictionary, name->dtype stype_dict : Dict of str->str Input storage type dictionary, name->storage_type group2ctx : Dict of string to mx.Context The dict mapping the `ctx_group` attribute to the context assignment. kwargs : Dict of str->shape Input shape dictionary, name->shape Returns ------- executor : mxnet.Executor An optimized TensorRT executor. def tensorrt_bind(symbol, ctx, all_params, type_dict=None, stype_dict=None, group2ctx=None, **kwargs): """Bind current symbol to get an optimized trt executor. Parameters ---------- symbol : Symbol The symbol you wish to bind, and optimize with TensorRT. ctx : Context The device context the generated executor to run on. all_params : Dict of str->ndarray A dictionary of mappings from parameter names to parameter NDArrays. type_dict : Dict of str->numpy.dtype Input type dictionary, name->dtype stype_dict : Dict of str->str Input storage type dictionary, name->storage_type group2ctx : Dict of string to mx.Context The dict mapping the `ctx_group` attribute to the context assignment. kwargs : Dict of str->shape Input shape dictionary, name->shape Returns ------- executor : mxnet.Executor An optimized TensorRT executor. """ kwargs['shared_buffer'] = all_params return symbol.simple_bind(ctx, type_dict=type_dict, stype_dict=stype_dict, group2ctx=group2ctx, **kwargs)
Parameters ---------- num_classes : int, default 1000 Number of classification classes. num_layers : int Number of layers for the variant of densenet. Options are 11, 13, 16, 19. batch_norm : bool, default False Use batch normalization. dtype: str, float32 or float16 Data precision. def get_symbol(num_classes, num_layers=11, batch_norm=False, dtype='float32', **kwargs): """ Parameters ---------- num_classes : int, default 1000 Number of classification classes. num_layers : int Number of layers for the variant of densenet. Options are 11, 13, 16, 19. batch_norm : bool, default False Use batch normalization. dtype: str, float32 or float16 Data precision. """ vgg_spec = {11: ([1, 1, 2, 2, 2], [64, 128, 256, 512, 512]), 13: ([2, 2, 2, 2, 2], [64, 128, 256, 512, 512]), 16: ([2, 2, 3, 3, 3], [64, 128, 256, 512, 512]), 19: ([2, 2, 4, 4, 4], [64, 128, 256, 512, 512])} if num_layers not in vgg_spec: raise ValueError("Invalide num_layers {}. Possible choices are 11,13,16,19.".format(num_layers)) layers, filters = vgg_spec[num_layers] data = mx.sym.Variable(name="data") if dtype == 'float16': data = mx.sym.Cast(data=data, dtype=np.float16) feature = get_feature(data, layers, filters, batch_norm) classifier = get_classifier(feature, num_classes) if dtype == 'float16': classifier = mx.sym.Cast(data=classifier, dtype=np.float32) symbol = mx.sym.SoftmaxOutput(data=classifier, name='softmax') return symbol
:param frame: an (w,h,channels) numpy array (image) :return: DataBatch of (1,channels,data_shape,data_shape) def create_batch(self, frame): """ :param frame: an (w,h,channels) numpy array (image) :return: DataBatch of (1,channels,data_shape,data_shape) """ frame_resize = mx.nd.array(cv2.resize(frame, (self.data_shape[0], self.data_shape[1]))) #frame_resize = mx.img.imresize(frame, self.data_shape[0], self.data_shape[1], cv2.INTER_LINEAR) # Change dimensions from (w,h,channels) to (channels, w, h) frame_t = mx.nd.transpose(frame_resize, axes=(2,0,1)) frame_norm = frame_t - self.mean_pixels_nd # Add dimension for batch, results in (1,channels,w,h) batch_frame = [mx.nd.expand_dims(frame_norm, axis=0)] batch_shape = [DataDesc('data', batch_frame[0].shape)] batch = DataBatch(data=batch_frame, provide_data=batch_shape) return batch
detect all images in iterator Parameters: ---------- det_iter : DetIter iterator for all testing images show_timer : Boolean whether to print out detection exec time Returns: ---------- list of detection results def detect_iter(self, det_iter, show_timer=False): """ detect all images in iterator Parameters: ---------- det_iter : DetIter iterator for all testing images show_timer : Boolean whether to print out detection exec time Returns: ---------- list of detection results """ num_images = det_iter._size if not isinstance(det_iter, mx.io.PrefetchingIter): det_iter = mx.io.PrefetchingIter(det_iter) start = timer() detections = self.mod.predict(det_iter).asnumpy() time_elapsed = timer() - start if show_timer: logging.info("Detection time for {} images: {:.4f} sec".format( num_images, time_elapsed)) result = Detector.filter_positive_detections(detections) return result
Return detections for batch :param batch: :return: def detect_batch(self, batch): """ Return detections for batch :param batch: :return: """ self.mod.forward(batch, is_train=False) detections = self.mod.get_outputs()[0] positive_detections = Detector.filter_positive_detections(detections) return positive_detections
wrapper for detecting multiple images Parameters: ---------- im_list : list of str image path or list of image paths root_dir : str directory of input images, optional if image path already has full directory information extension : str image extension, eg. ".jpg", optional Returns: ---------- list of detection results in format [det0, det1...], det is in format np.array([id, score, xmin, ymin, xmax, ymax]...) def im_detect(self, im_list, root_dir=None, extension=None, show_timer=False): """ wrapper for detecting multiple images Parameters: ---------- im_list : list of str image path or list of image paths root_dir : str directory of input images, optional if image path already has full directory information extension : str image extension, eg. ".jpg", optional Returns: ---------- list of detection results in format [det0, det1...], det is in format np.array([id, score, xmin, ymin, xmax, ymax]...) """ test_db = TestDB(im_list, root_dir=root_dir, extension=extension) test_iter = DetIter(test_db, 1, self.data_shape, self.mean_pixels, is_train=False) return self.detect_iter(test_iter, show_timer)
visualize detections in one image Parameters: ---------- img : numpy.array image, in bgr format dets : numpy.array ssd detections, numpy.array([[id, score, x1, y1, x2, y2]...]) each row is one object classes : tuple or list of str class names thresh : float score threshold def visualize_detection(self, img, dets, classes=[], thresh=0.6): """ visualize detections in one image Parameters: ---------- img : numpy.array image, in bgr format dets : numpy.array ssd detections, numpy.array([[id, score, x1, y1, x2, y2]...]) each row is one object classes : tuple or list of str class names thresh : float score threshold """ import matplotlib.pyplot as plt import random plt.imshow(img) height = img.shape[0] width = img.shape[1] colors = dict() for det in dets: (klass, score, x0, y0, x1, y1) = det if score < thresh: continue cls_id = int(klass) if cls_id not in colors: colors[cls_id] = (random.random(), random.random(), random.random()) xmin = int(x0 * width) ymin = int(y0 * height) xmax = int(x1 * width) ymax = int(y1 * height) rect = plt.Rectangle((xmin, ymin), xmax - xmin, ymax - ymin, fill=False, edgecolor=colors[cls_id], linewidth=3.5) plt.gca().add_patch(rect) class_name = str(cls_id) if classes and len(classes) > cls_id: class_name = classes[cls_id] plt.gca().text(xmin, ymin - 2, '{:s} {:.3f}'.format(class_name, score), bbox=dict(facecolor=colors[cls_id], alpha=0.5), fontsize=12, color='white') plt.show()
First column (class id) is -1 for negative detections :param detections: :return: def filter_positive_detections(detections): """ First column (class id) is -1 for negative detections :param detections: :return: """ class_idx = 0 assert(isinstance(detections, mx.nd.NDArray) or isinstance(detections, np.ndarray)) detections_per_image = [] # for each image for i in range(detections.shape[0]): result = [] det = detections[i, :, :] for obj in det: if obj[class_idx] >= 0: result.append(obj) detections_per_image.append(result) logging.info("%d positive detections", len(result)) return detections_per_image
wrapper for im_detect and visualize_detection Parameters: ---------- im_list : list of str or str image path or list of image paths root_dir : str or None directory of input images, optional if image path already has full directory information extension : str or None image extension, eg. ".jpg", optional Returns: ---------- def detect_and_visualize(self, im_list, root_dir=None, extension=None, classes=[], thresh=0.6, show_timer=False): """ wrapper for im_detect and visualize_detection Parameters: ---------- im_list : list of str or str image path or list of image paths root_dir : str or None directory of input images, optional if image path already has full directory information extension : str or None image extension, eg. ".jpg", optional Returns: ---------- """ dets = self.im_detect(im_list, root_dir, extension, show_timer=show_timer) if not isinstance(im_list, list): im_list = [im_list] assert len(dets) == len(im_list) for k, det in enumerate(dets): img = cv2.imread(im_list[k]) img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) self.visualize_detection(img, det, classes, thresh)
Runs the caffe upgrade tool on the prototxt to create a prototxt in the latest format. This enable us to work just with latest structures, instead of supporting all the variants :param caffe_root: link to caffe root folder, where the upgrade tool is located :param deploy_proto: name of the original prototxt file :return: name of new processed prototxt file def process_network_proto(caffe_root, deploy_proto): """ Runs the caffe upgrade tool on the prototxt to create a prototxt in the latest format. This enable us to work just with latest structures, instead of supporting all the variants :param caffe_root: link to caffe root folder, where the upgrade tool is located :param deploy_proto: name of the original prototxt file :return: name of new processed prototxt file """ processed_deploy_proto = deploy_proto + ".processed" from shutil import copyfile copyfile(deploy_proto, processed_deploy_proto) # run upgrade tool on new file name (same output file) import os upgrade_tool_command_line = caffe_root + '/build/tools/upgrade_net_proto_text.bin ' \ + processed_deploy_proto + ' ' + processed_deploy_proto os.system(upgrade_tool_command_line) return processed_deploy_proto
Reads from the caffe prototxt the network structure :param processed_deploy_prototxt: name of prototxt to load, preferably the prototxt should be processed before using a call to process_network_proto() :return: network_def, layer_name_to_record, top_to_layers network_def: caffe network structure, gives access to *all* the network information layer_name_to_record: *ordered* dictionary which maps between layer name and a structure which describes in a simple form the layer parameters top_to_layers: dictionary which maps a blob name to an ordered list of layers which output it when a top is used several times, like in inplace layhers, the list will contain all the layers by order of appearance def read_network_dag(processed_deploy_prototxt): """ Reads from the caffe prototxt the network structure :param processed_deploy_prototxt: name of prototxt to load, preferably the prototxt should be processed before using a call to process_network_proto() :return: network_def, layer_name_to_record, top_to_layers network_def: caffe network structure, gives access to *all* the network information layer_name_to_record: *ordered* dictionary which maps between layer name and a structure which describes in a simple form the layer parameters top_to_layers: dictionary which maps a blob name to an ordered list of layers which output it when a top is used several times, like in inplace layhers, the list will contain all the layers by order of appearance """ from caffe.proto import caffe_pb2 from google.protobuf import text_format # pylint: disable=relative-import from collections import OrderedDict # load prototxt file network_def = caffe_pb2.NetParameter() with open(processed_deploy_prototxt, 'r') as proto_file: text_format.Merge(str(proto_file.read()), network_def) # map layer name to layer record layer_name_to_record = OrderedDict() for layer_def in network_def.layer: if (len(layer_def.include) == 0) or \ (caffe_pb2.TEST in [item.phase for item in layer_def.include]): layer_name_to_record[layer_def.name] = LayerRecord(layer_def) top_to_layers = dict() for layer in network_def.layer: # no specific phase, or TEST phase is specifically asked for if (len(layer.include) == 0) or (caffe_pb2.TEST in [item.phase for item in layer.include]): for top in layer.top: if top not in top_to_layers: top_to_layers[top] = list() top_to_layers[top].append(layer.name) # find parents and children of all layers for child_layer_name in layer_name_to_record.keys(): # pylint: disable=too-many-nested-blocks child_layer_def = layer_name_to_record[child_layer_name] for bottom in child_layer_def.bottoms: if bottom in top_to_layers: for parent_layer_name in top_to_layers[bottom]: if parent_layer_name in layer_name_to_record: parent_layer_def = layer_name_to_record[parent_layer_name] if parent_layer_def not in child_layer_def.parents: child_layer_def.parents.append(parent_layer_def) if child_layer_def not in parent_layer_def.children: parent_layer_def.children.append(child_layer_def) # update filter, strid, pad for maxout "structures" for layer_name in layer_name_to_record.keys(): layer_def = layer_name_to_record[layer_name] if layer_def.type == 'Eltwise' and \ len(layer_def.parents) == 1 and \ layer_def.parents[0].type == 'Slice' and \ len(layer_def.parents[0].parents) == 1 and \ layer_def.parents[0].parents[0].type in ['Convolution', 'InnerProduct']: layer_def.filter = layer_def.parents[0].parents[0].filter layer_def.stride = layer_def.parents[0].parents[0].stride layer_def.pad = layer_def.parents[0].parents[0].pad return network_def, layer_name_to_record, top_to_layers