Skip to content

X
X2Paddle

PaddlePaddle / X2Paddle
大约 1 年 前同步成功

通知 328
Star 698
Fork 167
X
X2Paddle
提交 a6759829 编写于 作者: C Channingss
浏览文件
操作

Merge remote-tracking branch 'paddle/develop' into prior_box

上级 d93dc40c 78c614b7
隐藏空白更改
内联 并排
Showing with 245 addition and 278 deletion
op_list.md
浏览文件 @ a6759829
# X2Paddle支持OP列表 # X2Paddle支持OP列表
> 目前X2Paddle支持50+的TensorFlow OP，30+的Caffe Layer，覆盖了大部分CV分类模型常用的操作。我们在如下列表中给出了目前X2Paddle支持的全部OP。 > 目前X2Paddle支持70+的TensorFlow OP，30+的Caffe Layer，覆盖了大部分CV分类模型常用的操作。我们在如下列表中给出了目前X2Paddle支持的全部OP。
**注：** 目前，部分OP暂未支持，如您在转换过程中出现OP不支持的情况，可自行添加或反馈给我们。欢迎通过[ISSUE反馈](https://github.com/PaddlePaddle/X2Paddle/issues/new)的方式告知我们(模型名，代码实现或模型获取方式)，我们会及时跟进：） **注：** 目前，部分OP暂未支持，如您在转换过程中出现OP不支持的情况，可自行添加或反馈给我们。欢迎通过[ISSUE反馈](https://github.com/PaddlePaddle/X2Paddle/issues/new)的方式告知我们(模型名，代码实现或模型获取方式)，我们会及时跟进：）
...@@ -7,20 +7,24 @@ ...@@ -7,20 +7,24 @@
| 序号 | OP | 序号 | OP | 序号 | OP | 序号 | OP | | 序号 | OP | 序号 | OP | 序号 | OP | 序号 | OP |
|------|------|------|------|------|------|------|------| |------|------|------|------|------|------|------|------|
| 1 | Relu | 2 | Relu6 | 3 | Shape | 4 | Abs | | 1 | Relu | 2 | Relu6 | 3 | Shape | 4 | Abs |
| 5 | Sigmoid | 6 | Exp | 7 | Rsqrt | 8 | swish_f32 | | 5 | Sigmoid | 6 | Exp | 7 | Rsqrt | 8 | swish_f32 |
| 9 | Tanh | 10 | LeakyRelu | 11 | Add | 12 | RealDiv | | 9 | Tanh | 10 | LeakyRelu | 11 | Add | 12 | RealDiv |
| 13 | Sub | 14 | Maximum | 15 | Mul | 16 | FloorDiv | | 13 | Sub | 14 | Maximum | 15 | Mul | 16 | FloorDiv |
| 17 | Placeholder | 18 | Const | 19 | Transpose | 20 | FusedBatchNorm | | 17 | Placeholder | 18 | Const | 19 | Transpose | 20 | FusedBatchNorm |
| 21 | Conv2D | 22 | BiasAdd | 23 | MaxPool | 24 | DepthwiseConv2dNative | | 21 | Conv2D | 22 | BiasAdd | 23 | MaxPool | 24 | DepthwiseConv2dNative |
| 25 | Reshape | 26 | AvgPool | 27 | SplitV | 28 | SquaredDifference | | 25 | Reshape | 26 | AvgPool | 27 | SplitV | 28 | SquaredDifference |
| 29 | Tile | 30 | Pack | 31 | Pad | 32 | ResizeBilinear | | 29 | Tile | 30 | Pack | 31 | Pad | 32 | ResizeBilinear |
| 33 | Mean | 34 | MatMul | 35 | ArgMax | 36 | StridedSlice | | 33 | Mean | 34 | MatMul | 35 | ArgMax | 36 | StridedSlice |
| 37 | Slice | 38 | Sum | 39 | Max | 40 | Conv2DBackpropInput | | 37 | Slice | 38 | Sum | 39 | Max | 40 | Conv2DBackpropInput |
| 41 | Cast | 42 | Split | 43 | Squeeze | 44 | ResizeNearestNeighbor | | 41 | Cast | 42 | Split | 43 | Squeeze | 44 | ResizeNearestNeighbor |
| 45 | Softmax | 46 | Range | 47 | ConcatV2 | 48 | MirrorPad | | 45 | Softmax | 46 | Range | 47 | ConcatV2 | 48 | MirrorPad |
| 49 | Identity | 50 | GreaterEqual | 51 | StopGradient | 52 | Minimum | | 49 | Identity | 50 | GreaterEqual | 51 | StopGradient | 52 | Minimum |
| 53 | RadnomUniform | 54 | Fill | 55 | Floor | 56 | DepthToSpace | | 53 | RadnomUniform | 54 | Fill | 55 | Floor | 56 | DepthToSpace |
| 57 | Sqrt | 58 | Softplus | 59 | Erf | 60 | AddV2 |
| 61 | LessEqual | 62 | BatchMatMul | 63 | BatchMatMulV2 | 64 | ExpandDims |
| 65 | BatchToSpaceND | 66 | SpaceToBatchND | 67 | OneHot | 68 | Pow |
| 69 | All | 70 | GatherV2 | 71 | IteratorV2 | | |
## Caffe ## Caffe
......
x2paddle/decoder/onnx_shape_inference.py
浏览文件 @ a6759829
...@@ -267,9 +267,8 @@ class SymbolicShapeInference: ...@@ -267,9 +267,8 @@ class SymbolicShapeInference:
if pending_nodes and self.verbose_ > 0: if pending_nodes and self.verbose_ > 0:
print('SymbolicShapeInference: orphaned nodes discarded: ') print('SymbolicShapeInference: orphaned nodes discarded: ')
print( print('\n'.join(
* [n.op_type + ': ' + n.output[0] for n in pending_nodes], [n.op_type + ': ' + n.output[0] for n in pending_nodes]))
sep='\n')
if input_shapes is not None: if input_shapes is not None:
for input_name, shape in input_shapes.items(): for input_name, shape in input_shapes.items():
for idx in range(len(self.out_mp_.graph.input)): for idx in range(len(self.out_mp_.graph.input)):
......
x2paddle/op_mapper/caffe_custom_layer/normalize.py
浏览文件 @ a6759829
...@@ -17,7 +17,7 @@ def normalize_layer(inputs, ...@@ -17,7 +17,7 @@ def normalize_layer(inputs,
scale_param = fluid.layers.create_parameter( scale_param = fluid.layers.create_parameter(
shape=[1] if channel_shared else [1, 1, 1, input_shape[0][1]], shape=[1] if channel_shared else [1, 1, 1, input_shape[0][1]],
dtype=input.dtype, dtype=input.dtype,
attr=name + '_scale') attr=fluid.ParamAttr(name=name + '_scale'))
scale_param = fluid.layers.reshape(x=scale_param, \ scale_param = fluid.layers.reshape(x=scale_param, \
shape=[1] if channel_shared else [input_shape[0][1]]) shape=[1] if channel_shared else [input_shape[0][1]])
out = fluid.layers.elementwise_mul( out = fluid.layers.elementwise_mul(
......
x2paddle/op_mapper/onnx2paddle/opset9/opset.py
浏览文件 @ a6759829
...@@ -32,15 +32,33 @@ import shutil ...@@ -32,15 +32,33 @@ import shutil
_logger = _logging.getLogger(__name__) _logger = _logging.getLogger(__name__)
def _const_weight_or_none(node): def _const_weight_or_none(node, necessary=False):
if 'Constant' in node.layer_type: if 'Constant' in node.layer_type:
return node.value return node.value
if isinstance(node, ONNXGraphDataNode): if isinstance(node, ONNXGraphDataNode):
return node.weight return node.weight
if necessary:
assert '{} should be an initializer or Constant operator.'.format(
node.layer_name)
return None return None
def get_same_padding(in_size, kernel_size, stride): def _is_static_shape(shape):
negtive_dims = 0
error_dims = 0
for dim in shape:
if dim < 0:
negtive_dims += 1
if dim < -1:
error_dims += 1
if negtive_dims > 1:
return False
if error_dims > 0:
return False
return True
def _get_same_padding(in_size, kernel_size, stride):
new_size = int(math.ceil(in_size * 1.0 / stride)) new_size = int(math.ceil(in_size * 1.0 / stride))
pad_size = (new_size - 1) * stride + kernel_size - in_size pad_size = (new_size - 1) * stride + kernel_size - in_size
pad0 = int(pad_size / 2) pad0 = int(pad_size / 2)
...@@ -228,42 +246,9 @@ class OpSet9(): ...@@ -228,42 +246,9 @@ class OpSet9():
val_x = self.graph.get_input_node(node, idx=0, copy=True) val_x = self.graph.get_input_node(node, idx=0, copy=True)
val_y = self.graph.get_input_node(node, idx=1, copy=True) val_y = self.graph.get_input_node(node, idx=1, copy=True)
val_y_shape = val_y.out_shapes[0] inputs = {'x': val_x, 'y': val_y}
val_x_shape = val_x.out_shapes[0] node.fluid_code.add_layer(
op_type, inputs=inputs, output=node, param_attr=None)
if len(val_x_shape) < len(val_y_shape):
val_x, val_y = val_y, val_x
val_y_shape, val_x_shape = val_x_shape, val_y_shape
str_y_shape = ','.join(str(e) for e in val_y_shape)
str_x_shape = ','.join(str(e) for e in val_x_shape)
slice_idx = 0
if str_y_shape not in str_x_shape:
for dim in val_y_shape:
if dim == 1:
slice_idx += 1
else:
break
attr = {"name": string(node.layer_name)}
if slice_idx < len(val_y_shape) and slice_idx > 0:
val_y_reshaped = val_y_shape[slice_idx:]
var_y_reshaped = val_y.layer_name + '_reshaped'
attr_reshaped = {
'shape': val_y_reshaped,
'name': string(var_y_reshaped)
}
node.fluid_code.add_layer(
'reshape',
inputs=val_y,
output=var_y_reshaped,
param_attr=attr_reshaped)
inputs = {'x': val_x, 'y': var_y_reshaped}
node.fluid_code.add_layer(
op_type, inputs=inputs, output=node, param_attr=attr)
else:
inputs = {'x': val_x, 'y': val_y}
node.fluid_code.add_layer(
op_type, inputs=inputs, output=node, param_attr=attr)
@print_mapping_info @print_mapping_info
def place_holder(self, node): def place_holder(self, node):
...@@ -476,8 +461,21 @@ class OpSet9(): ...@@ -476,8 +461,21 @@ class OpSet9():
output=node, output=node,
param_attr={'shape': [1]}) param_attr={'shape': [1]})
else: else:
node.fluid_code.add_layer( if str(val_x.dtype) == 'bool':
'unsqueeze', inputs=val_x, output=node, param_attr=attr) val_x_cast = val_x.layer_name + '_cast'
node.fluid_code.add_layer(
'cast',
inputs=val_x,
output=val_x_cast,
param_attr={'dtype': string('int64')})
node.fluid_code.add_layer(
'unsqueeze',
inputs=val_x_cast,
output=node,
param_attr=attr)
else:
node.fluid_code.add_layer(
'unsqueeze', inputs=val_x, output=node, param_attr=attr)
@print_mapping_info @print_mapping_info
def Shrink(self, node): def Shrink(self, node):
...@@ -597,12 +595,35 @@ class OpSet9(): ...@@ -597,12 +595,35 @@ class OpSet9():
#assert len( #assert len(
# indices_shape) <= 2, "Gather op don't support dim of indice >2 " # indices_shape) <= 2, "Gather op don't support dim of indice >2 "
if axis == 0 and len(indices_shape) <= 1: if axis == 0 and len(indices_shape) <= 1:
node.fluid_code.add_layer( if len(val_x.out_shapes[0]) <= 1:
'gather', node.fluid_code.add_layer(
inputs={'input': val_x, 'gather',
'index': indices}, inputs={'input': val_x,
output=node, 'index': indices},
param_attr=None) output=node,
param_attr=None)
elif len(val_x.out_shapes[0]) > 1:
if len(indices_shape) == 0:
gather_ = node.layer_name + '_1'
node.fluid_code.add_layer(
'gather',
inputs={'input': val_x,
'index': indices},
output=gather_,
param_attr=None)
node.fluid_code.add_layer(
'squeeze',
inputs={'input': gather_,
'axes': [0]},
output=node,
param_attr=None)
else:
node.fluid_code.add_layer(
'gather',
inputs={'input': val_x,
'index': indices},
output=node,
param_attr=None)
elif axis > 0 and len(indices_shape) <= 1: elif axis > 0 and len(indices_shape) <= 1:
perm = list(range(len(val_x.out_shapes[0]))) perm = list(range(len(val_x.out_shapes[0])))
perm = [axis] + perm[:axis] + perm[axis + 1:] perm = [axis] + perm[:axis] + perm[axis + 1:]
...@@ -621,6 +642,13 @@ class OpSet9(): ...@@ -621,6 +642,13 @@ class OpSet9():
param_attr=None) param_attr=None)
node.fluid_code.add_layer( node.fluid_code.add_layer(
'transpose', inputs=node, output=node, param_attr=attr_trans) 'transpose', inputs=node, output=node, param_attr=attr_trans)
if len(indices_shape) < 1:
node.fluid_code.add_layer(
'squeeze',
inputs={'input': node,
'axes': [0]},
output=node,
param_attr=None)
elif axis == 0 and len(indices_shape) > 1: elif axis == 0 and len(indices_shape) > 1:
if val_x.out_shapes[0] is not None and isinstance( if val_x.out_shapes[0] is not None and isinstance(
val_x, ONNXGraphDataNode): val_x, ONNXGraphDataNode):
...@@ -701,6 +729,86 @@ class OpSet9(): ...@@ -701,6 +729,86 @@ class OpSet9():
output=node, output=node,
param_attr={'shape': reshaped_shape}) param_attr={'shape': reshaped_shape})
@print_mapping_info
def ScatterND(self, node):
val_x = self.graph.get_input_node(node, idx=0, copy=True)
indices = self.graph.get_input_node(node, idx=1, copy=True)
updates = self.graph.get_input_node(node, idx=2, copy=True)
if len(indices.out_shapes[0]) == 1:
node.fluid_code.add_layer(
'scatter',
inputs={'input': val_x,
'index': indices,
'updates': updates},
output=node,
param_attr=None)
else:
input_inner_indices = node.layer_name + '_input_inner_indices'
node.fluid_code.add_layer(
'scatter_nd',
inputs={
'shape': val_x.out_shapes[0],
'index': indices,
'updates': updates
},
output=input_inner_indices,
param_attr=None)
constant_minus_one = node.layer_name + '_constant_minus_one'
node.fluid_code.add_layer(
'fill_constant',
inputs=None,
output=constant_minus_one,
param_attr={
'shape': updates.out_shapes[0],
'dtype': string(updates.dtype),
'value': -1
})
indices_mask = node.layer_name + '_indices_mask'
node.fluid_code.add_layer(
'scatter_nd',
inputs={
'shape': val_x.out_shapes[0],
'index': indices,
'updates': constant_minus_one
},
output=indices_mask,
param_attr=None)
constant_1 = node.layer_name + '_constant_1'
node.fluid_code.add_layer(
'fill_constant',
inputs=None,
output=constant_1,
param_attr={
'shape': val_x.out_shapes[0],
'dtype': string(val_x.dtype),
'value': 1
})
input_out_indices_mask = node.layer_name + '_input_out_indices_mask'
node.fluid_code.add_layer(
"elementwise_add",
inputs={"x": indices_mask,
"y": constant_1},
output=input_out_indices_mask,
param_attr=None)
input_out_indices = node.layer_name + '_input_out_indices'
node.fluid_code.add_layer(
"elementwise_mul",
inputs={"x": val_x,
"y": input_out_indices_mask},
output=input_out_indices,
param_attr=None)
node.fluid_code.add_layer(
"elementwise_add",
inputs={"x": input_inner_indices,
"y": input_out_indices},
output=node,
param_attr=None)
@print_mapping_info @print_mapping_info
def Range(self, node): def Range(self, node):
val_start = self.graph.get_input_node(node, idx=0, copy=True) val_start = self.graph.get_input_node(node, idx=0, copy=True)
...@@ -724,7 +832,7 @@ class OpSet9(): ...@@ -724,7 +832,7 @@ class OpSet9():
ends = self.graph.get_input_node(node, idx=2, copy=True) ends = self.graph.get_input_node(node, idx=2, copy=True)
if len(node.inputs) > 3: if len(node.inputs) > 3:
axes = self.graph.get_input_node(node, idx=3, copy=True) axes = self.graph.get_input_node(node, idx=3, copy=True)
axes = _const_weight_or_none(axes) axes = _const_weight_or_none(axes, necessary=True)
if len(node.inputs) > 4: if len(node.inputs) > 4:
steps = self.graph.get_input_node(node, idx=4, copy=True) steps = self.graph.get_input_node(node, idx=4, copy=True)
steps = _const_weight_or_none(steps) steps = _const_weight_or_none(steps)
...@@ -828,6 +936,14 @@ class OpSet9(): ...@@ -828,6 +936,14 @@ class OpSet9():
inputs={'x': val_x}, inputs={'x': val_x},
output=node, output=node,
param_attr={'shape': shape_value.tolist()}) param_attr={'shape': shape_value.tolist()})
elif len(node.out_shapes[0]) > 0 and _is_static_shape(node.out_shapes[
0]):
node.fluid_code.add_layer(
'reshape',
inputs={'x': val_x,
'shape': node.out_shapes[0]},
output=node,
param_attr=attr)
elif val_shape.dtype == 'int64': elif val_shape.dtype == 'int64':
val_shape_cast = val_shape.layer_name + '_cast' val_shape_cast = val_shape.layer_name + '_cast'
node.fluid_code.add_layer( node.fluid_code.add_layer(
...@@ -879,6 +995,11 @@ class OpSet9(): ...@@ -879,6 +995,11 @@ class OpSet9():
node.fluid_code.add_layer( node.fluid_code.add_layer(
'cast', inputs=val_input, output=node, param_attr=attr) 'cast', inputs=val_input, output=node, param_attr=attr)
@print_mapping_info
def Not(self, node):
val_input = self.graph.get_input_node(node, idx=0, copy=True)
node.fluid_code.add_layer('logical_not', inputs=val_input, output=node)
@print_mapping_info @print_mapping_info
def AveragePool(self, node): def AveragePool(self, node):
val_x = self.graph.get_input_node(node, idx=0, copy=True) val_x = self.graph.get_input_node(node, idx=0, copy=True)
...@@ -897,11 +1018,11 @@ class OpSet9(): ...@@ -897,11 +1018,11 @@ class OpSet9():
if auto_pad == "SAME_UPPER" or auto_pad == "SAME_LOWER": if auto_pad == "SAME_UPPER" or auto_pad == "SAME_LOWER":
input_shape = val_x.out_shapes[0] input_shape = val_x.out_shapes[0]
pad_h = get_same_padding(input_shape[2], kernel_shape[0], pad_h = _get_same_padding(input_shape[2], kernel_shape[0],
strides[0]) strides[0])
pad_w = get_same_padding(input_shape[3], kernel_shape[1], pad_w = _get_same_padding(input_shape[3], kernel_shape[1],
strides[1]) strides[1])
attr = {"paddings": pad_h + pad_w, "pad_value": 0.0} paddings = pad_h + pad_w
attr = { attr = {
"pool_size": kernel_shape, "pool_size": kernel_shape,
...@@ -1171,7 +1292,6 @@ class OpSet9(): ...@@ -1171,7 +1292,6 @@ class OpSet9():
def NonZero(self, node): def NonZero(self, node):
val_x = self.graph.get_input_node(node, idx=0, copy=True) val_x = self.graph.get_input_node(node, idx=0, copy=True)
val_x_dim = len(val_x.out_shapes[0]) val_x_dim = len(val_x.out_shapes[0])
print(val_x.layer_name, val_x.out_shapes[0])
if val_x_dim == 1: if val_x_dim == 1:
node.fluid_code.add_layer("nonzero", inputs=val_x, output=val_x) node.fluid_code.add_layer("nonzero", inputs=val_x, output=val_x)
node.fluid_code.add_layer( node.fluid_code.add_layer(
...@@ -1232,11 +1352,11 @@ class OpSet9(): ...@@ -1232,11 +1352,11 @@ class OpSet9():
if auto_pad == "SAME_UPPER" or auto_pad == "SAME_LOWER": if auto_pad == "SAME_UPPER" or auto_pad == "SAME_LOWER":
input_shape = val_x.out_shapes[0] input_shape = val_x.out_shapes[0]
pad_h = get_same_padding(input_shape[2], kernel_shape[0], pad_h = _get_same_padding(input_shape[2], kernel_shape[0],
strides[0]) strides[0])
pad_w = get_same_padding(input_shape[3], kernel_shape[1], pad_w = _get_same_padding(input_shape[3], kernel_shape[1],
strides[1]) strides[1])
attr = {"paddings": pad_h + pad_w, "pad_value": 0.0} paddings = pad_h + pad_w
attr = { attr = {
"pool_size": kernel_shape, "pool_size": kernel_shape,
...@@ -1293,23 +1413,23 @@ class OpSet9(): ...@@ -1293,23 +1413,23 @@ class OpSet9():
kernel_shape = node.get_attr('kernel_shape') kernel_shape = node.get_attr('kernel_shape')
convnd = len(kernel_shape) convnd = len(kernel_shape)
assert 2 <= convnd <= 3, 'only conv2d and conv3d is supported' assert 2 <= convnd <= 3, 'only conv2d and conv3d is supported'
num_out_channels = val_w.out_shapes[0][0] # OI... num_out_channels = val_w.out_shapes[0][0]
fluid_op = 'conv{}d'.format(convnd) fluid_op = 'conv{}d'.format(convnd)
num_groups = node.get_attr('group', 1) num_groups = node.get_attr('group', 1)
strides = node.get_attr('strides', [1] * convnd) # optional strides = node.get_attr('strides', [1] * convnd)
dilations = node.get_attr('dilations', [1] * convnd) # optional dilations = node.get_attr('dilations', [1] * convnd)
pads = node.get_attr('pads', [0] * (convnd * 2)) # optional pads = node.get_attr('pads', [0] * (convnd * 2))
input_shape = val_x.out_shapes[0] input_shape = val_x.out_shapes[0]
paddings, val_x = self._pad_if_asymmetric(node, pads, val_x) paddings, val_x = self._pad_if_asymmetric(node, pads, val_x)
if auto_pad == "SAME_UPPER" or auto_pad == "SAME_LOWER": if auto_pad == "SAME_UPPER" or auto_pad == "SAME_LOWER":
pad_h = get_same_padding(input_shape[2], kernel_shape[0], pad_h = _get_same_padding(input_shape[2], kernel_shape[0],
strides[0]) strides[0])
pad_w = get_same_padding(input_shape[3], kernel_shape[1], pad_w = _get_same_padding(input_shape[3], kernel_shape[1],
strides[1]) strides[1])
attr = {"paddings": pad_h + pad_w, "pad_value": 0.0} paddings = pad_h + pad_w
attr = { attr = {
"num_filters": num_out_channels, "num_filters": num_out_channels,
...@@ -1379,183 +1499,3 @@ class OpSet9(): ...@@ -1379,183 +1499,3 @@ class OpSet9():
} }
node.fluid_code.add_layer( node.fluid_code.add_layer(
fluid_op, inputs=val_x, output=node, param_attr=attr) fluid_op, inputs=val_x, output=node, param_attr=attr)
@print_mapping_info
def GRU(self, node):
val_x = self.graph.get_input_node(node, idx=0, copy=True)
val_w = self.graph.get_input_node(node, idx=1, copy=True)
val_r = self.graph.get_input_node(node, idx=2, copy=True)
val_b = None
val_len = None
val_xh = None
miss_arg_num = 0
num_ipt = len(node.layer.input)
if num_ipt > 3 and node.layer.input[3] != '':
val_b = self.graph.get_input_node(node, idx=3, copy=True)
else:
miss_arg_num += 1
if num_ipt > 4 and node.layer.input[4] != '':
val_len = self.graph.get_input_node(
node, idx=4 - miss_arg_num, copy=True)
else:
miss_arg_num += 1
if num_ipt > 5 and node.layer.input[5] != '':
val_xh = self.graph.get_input_node(
node, idx=5 - miss_arg_num, copy=True)
x_shape = val_x.out_shapes[0]
assert x_shape[1] == 1, 'only X with batch_size = 1 supported'
assert node.get_attr('clip', None) is None, 'clipping not supported'
hidden_size = node.get_attr('hidden_size', None)
if hidden_size is None:
r_shape = val_r.out_shapes[0]
if r_shape:
hidden_size = r_shape[-1]
if hidden_size is None:
w_shape = var_w.out_shapes[0]
if w_shape:
hidden_size = w_shape[-2] // 3
if hidden_size is None and val_b:
b_shape = val_b.out_shapes[0]
if b_shape:
hidden_size = b_shape[-1] // 6
if hidden_size is None and val_xh:
xh_shape = val_xh.out_shapes[0]
if xh_shape:
hidden_size = xh_shape[-1]
direction = node.get_attr('direction', 'forward')
assert direction != 'bidirectional', 'direction = bidirectional not supported'
activations = node.get_attr('activations', ['Sigmoid', 'Tanh'])
assert len(activations) == 2, 'bidirectional operation not supported'
assert node.get_attr('linear_before_reset',
0) == 0, 'only linear_before_reset = 0 supported'
activations = [s.lower() for s in activations]
gate_activation, candidate_activation = activations
is_reverse = direction == 'reverse'
var_x0 = node.layer_name + '_x0'
node.fluid_code.add_layer(
'squeeze',
inputs=val_x,
output=var_x0,
param_attr={'axes': [1],
'name': string(var_x0)})
var_w0 = node.layer_name + '_w0'
node.fluid_code.add_layer(
'squeeze',
inputs=val_w,
output=var_w0,
param_attr={'axes': [0],
'name': string(var_w0)})
var_fc = node.layer_name + '_fc'
var_mm = (node.layer_name + '_mm') if val_b else var_fc
node.fluid_code.add_layer(
'matmul',
inputs={'x': var_x0,
'y': var_w0},
output=var_mm,
param_attr={
'transpose_x': 0,
'transpose_y': 1,
'name': string(var_mm)
})
var_r0 = node.layer_name + '_r0'
node.fluid_code.add_layer(
'squeeze',
inputs=val_r,
output=var_r0,
param_attr={'axes': [0],
'name': string(var_r0)})
var_r0t = node.layer_name + '_r0t'
node.fluid_code.add_layer(
'transpose',
inputs=var_r0,
output=var_r0t,
param_attr={'perm': [1, 0],
'name': string(var_r0t)})
if val_b:
var_bi = node.layer_name + '_bi'
var_bh = node.layer_name + '_bh'
node.fluid_code.add_layer(
'split',
inputs=val_b,
output=var_bi + ',' + var_bh,
param_attr={
'dim': 1,
'num_or_sections': [hidden_size * 3, hidden_size * 3],
'name': string(node.layer_name + '.b/split')
})
var_bi0 = node.layer_name + '_bi0'
node.fluid_code.add_layer(
'squeeze',
inputs=var_bi,
output=var_bi0,
param_attr={'axes': [0],
'name': string(var_bi0)})
node.fluid_code.add_layer(
'elementwise_add',
inputs=[var_mm, var_bi0],
output=var_fc,
param_attr={
'axes': 1,
'name': string(node.layer_name + '.i/bias')
})
if val_xh:
var_xh0 = node.layer_name + '_xh0'
node.fluid_code.add_layer(
'squeeze',
inputs=val_xh,
output=var_xh0,
param_attr={'axes': [1],
'name': string(var_xh0)})
var_y00 = node.layer_name + '_y00'
attr = {
'origin_mode': True,
'h_0': var_xh0 if val_xh else None,
'is_reverse': is_reverse,
'gate_activation': string(gate_activation),
'candidate_activation': string(candidate_activation),
'param_attr': string(var_r0t),
'bias_attr': string(var_bh) if val_b else False,
}
node.fluid_code.add_layer(
'dynamic_gru',
inputs=var_fc + ',' + str(hidden_size),
output=var_y00,
param_attr=attr)
num_opt = len(node.layer.output)
if num_opt > 0 and node.layer.output[0] != '':
node.fluid_code.add_layer(
'unsqueeze',
inputs=var_y00,
output=node.layer.output[0],
param_attr={
'axes': [1, 1],
'name': string(node.layer.output[0])
})
if num_opt > 1 and node.layer.output[1] != '':
node.fluid_code.add_layer(
'unsqueeze',
inputs=var_y00,
output=node.layer.output[1],
param_attr={
'axes': [1, 1],
'name': string(node.layer.output[1])
})
x2paddle/op_mapper/paddle2onnx/opset9/opset.py
浏览文件 @ a6759829
...@@ -875,6 +875,14 @@ class OpSet9(object): ...@@ -875,6 +875,14 @@ class OpSet9(object):
axes=op.attr('axes')) axes=op.attr('axes'))
return node return node
def cast(self, op, block):
node = helper.make_node(
'Cast',
inputs=op.input('X'),
outputs=op.output('Out'),
to=self.paddle_onnx_dtype_map[op.attr('out_dtype')])
return node
def arg_max(self, op, block): def arg_max(self, op, block):
node = helper.make_node( node = helper.make_node(
'ArgMax', 'ArgMax',
......
x2paddle/op_mapper/tf_op_mapper_nhwc.py
浏览文件 @ a6759829
...@@ -299,6 +299,10 @@ class TFOpMapperNHWC(OpMapper): ...@@ -299,6 +299,10 @@ class TFOpMapperNHWC(OpMapper):
data_format = node.get_attr("data_format").decode() data_format = node.get_attr("data_format").decode()
pad_mode = node.get_attr("padding").decode() pad_mode = node.get_attr("padding").decode()
channel_first = data_format == "NCHW" channel_first = data_format == "NCHW"
if data_format == "NHWC":
n, h, w, c = input.out_shapes[0]
else:
n, c, h, w = input.out_shapes[0]
if kernel.layer_type == 'Const': if kernel.layer_type == 'Const':
kernel_value = kernel.value kernel_value = kernel.value
...@@ -329,10 +333,15 @@ class TFOpMapperNHWC(OpMapper): ...@@ -329,10 +333,15 @@ class TFOpMapperNHWC(OpMapper):
"dilation": dilations[2:4], "dilation": dilations[2:4],
"padding": string(pad_mode) "padding": string(pad_mode)
} }
if hasattr(node, 'dilation') and attr['dilation'] == [1, 1]: if hasattr(node, 'dilation') and attr['dilation'] == [1, 1]:
if len(node.dilation) == 1: if len(node.dilation) == 1:
attr['dilation'] = [1, node.dilation[0]] attr['dilation'] = [1, node.dilation[0]]
if c == -1:
reshape_attr = {"shape": [0, k_size[2], 0, 0]}
node.fluid_code.add_layer(
"reshape", inputs=input, output=input, param_attr=reshape_attr)
node.fluid_code.add_layer( node.fluid_code.add_layer(
"conv2d", inputs=input, output=node, param_attr=attr) "conv2d", inputs=input, output=node, param_attr=attr)
if not channel_first: if not channel_first:
...@@ -748,11 +757,12 @@ class TFOpMapperNHWC(OpMapper): ...@@ -748,11 +757,12 @@ class TFOpMapperNHWC(OpMapper):
self.add_omit_nodes(begin.layer_name, node.layer_name) self.add_omit_nodes(begin.layer_name, node.layer_name)
begin = begin.value.tolist() begin = begin.value.tolist()
else: else:
begin = begin begin = self.decoder.infer_tensor(begin).tolist()
shape = begin.out_shapes[0]
attr = {"shape": shape} # shape = begin.out_shapes[0]
node.fluid_code.add_layer( # attr = {"shape": shape}
"reshape", inputs=begin, output=begin, param_attr=attr) # node.fluid_code.add_layer(
# "reshape", inputs=begin, output=begin, param_attr=attr)
if size.layer_type == "Const": if size.layer_type == "Const":
self.add_omit_nodes(size.layer_name, node.layer_name) self.add_omit_nodes(size.layer_name, node.layer_name)
size = size.value.tolist() size = size.value.tolist()
......
x2paddle/optimizer/tf_optimizer.py
浏览文件 @ a6759829
...@@ -863,6 +863,9 @@ class TFOptimizer(object): ...@@ -863,6 +863,9 @@ class TFOptimizer(object):
weight = numpy.expand_dims(weight, 2) weight = numpy.expand_dims(weight, 2)
weight = numpy.expand_dims(weight, 3) weight = numpy.expand_dims(weight, 3)
self.op_mapper.weights[in_nodes3[0].layer_name] = weight self.op_mapper.weights[in_nodes3[0].layer_name] = weight
# fix bug in Paddle1.8.3 and may change in next version.
self.op_mapper.weights[in_nodes3[0].layer_name +
'_1'] = weight.reshape(1, -1)
in_nodes3[0].fluid_code.layers[0].param_attr["shape"] = [ in_nodes3[0].fluid_code.layers[0].param_attr["shape"] = [
1, in_shape[-1], 1, 1 1, in_shape[-1], 1, 1
] ]
...@@ -885,7 +888,7 @@ class TFOptimizer(object): ...@@ -885,7 +888,7 @@ class TFOptimizer(object):
node.fluid_code.clear() node.fluid_code.clear()
attr = { attr = {
"mode": string(mode), "mode": string(mode),
"param_attr": string(in_nodes3[0].layer_name) "param_attr": string(in_nodes3[0].layer_name + "_1")
} }
node.fluid_code.add_layer( node.fluid_code.add_layer(
......
x2paddle_model_zoo.md
浏览文件 @ a6759829
# X2Paddle模型测试库 # X2Paddle模型测试库
> 目前X2Paddle支持50+的TensorFlow OP，40+的Caffe Layer，覆盖了大部分CV分类模型常用的操作。我们在如下模型列表中测试了X2Paddle的转换。 > 目前X2Paddle支持70+的TensorFlow OP，40+的Caffe Layer，覆盖了大部分CV分类模型常用的操作。我们在如下模型列表中测试了X2Paddle的转换。
**注：** 受限于不同框架的差异，部分模型可能会存在目前无法转换的情况，如TensorFlow中包含控制流的模型，NLP模型等。对于CV常见的模型，如若您发现无法转换或转换失败，存在较大diff等问题，欢迎通过[ISSUE反馈](https://github.com/PaddlePaddle/X2Paddle/issues/new)的方式告知我们(模型名，代码实现或模型获取方式)，我们会及时跟进：） **注：** 受限于不同框架的差异，部分模型可能会存在目前无法转换的情况，如TensorFlow中包含控制流的模型，NLP模型等。对于CV常见的模型，如若您发现无法转换或转换失败，存在较大diff等问题，欢迎通过[ISSUE反馈](https://github.com/PaddlePaddle/X2Paddle/issues/new)的方式告知我们(模型名，代码实现或模型获取方式)，我们会及时跟进：）
...@@ -20,10 +20,13 @@ ...@@ -20,10 +20,13 @@
| ResNet_V1_101 | [code](https://github.com/tensorflow/models/tree/master/research/slim/nets) |-| | ResNet_V1_101 | [code](https://github.com/tensorflow/models/tree/master/research/slim/nets) |-|
| ResNet_V2_101 | [code](https://github.com/tensorflow/models/tree/master/research/slim/nets) |-| | ResNet_V2_101 | [code](https://github.com/tensorflow/models/tree/master/research/slim/nets) |-|
| UNet | [code1](https://github.com/jakeret/tf_unet )/[code2](https://github.com/lyatdawn/Unet-Tensorflow) |-| | UNet | [code1](https://github.com/jakeret/tf_unet )/[code2](https://github.com/lyatdawn/Unet-Tensorflow) |-|
|MTCNN | [code](https://github.com/AITTSMD/MTCNN-Tensorflow) |-| | MTCNN | [code](https://github.com/AITTSMD/MTCNN-Tensorflow) |-|
|YOLO-V3| [code](https://github.com/YunYang1994/tensorflow-yolov3) | 转换需要关闭NHWC->NCHW的优化，见[文档Q2](FAQ.md) | | YOLO-V3| [code](https://github.com/YunYang1994/tensorflow-yolov3) | 转换需要关闭NHWC->NCHW的优化，见[文档Q2](FAQ.md) |
| FALSR | [code](https://github.com/xiaomi-automl/FALSR) | - | | FALSR | [code](https://github.com/xiaomi-automl/FALSR) | 需使用参数without_data_format_optimization |
| DCSCN | [code](https://modelzoo.co/model/dcscn-super-resolution) | - | | DCSCN | [code](https://modelzoo.co/model/dcscn-super-resolution) | 需使用参数without_data_format_optimization |
| Bert（albert） | [code](https://github.com/google-research/albert#pre-trained-models) | 需使用参数without_data_format_optimization |
| Bert（chinese_L-12_H-768_A-12） | [code](https://github.com/google-research/bert#pre-trained-models) | 需使用参数without_data_format_optimization |
| Bert（multi_cased_L-12_H-768_A-12） | [code](https://github.com/google-research/bert#pre-trained-models) | 需使用参数without_data_format_optimization |
## Caffe ## Caffe
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或