def compile_model(model, inputs) -> 'ONNXModel':
inputs_, outputs_, graph_ = core.convert_model(model, inputs)
if graph_ is None:
return None
preprocess(graph_, True)
generator = ONNXGenerator()
model = generator.generate_model(graph_.input_values, graph_.output_values,
graph_)
# check inputs
onnx_model = ONNXModel()
onnx_model.model = model
onnx_model.inputs = graph_.input_values
onnx_model.outputs = graph_.output_values
return onnx_model
def compile_model(model, inputs) -> 'ONNXModel':
# assign names
assigned_names.clear()
node2onnx_parameter.clear()
value2onnx_parameter.clear()
inputs_, outputs_, graph_ = core.convert_model(model, inputs)
if graph_ is None:
return None
preprocess(graph_, True)
generator = ONNXGenerator()
model = generator.generate_model(graph_.input_values, graph_.output_values,
graph_)
# check inputs
onnx_model = ONNXModel()
onnx_model.model = model
onnx_model.inputs = graph_.input_values
onnx_model.outputs = graph_.output_values
return onnx_model
def compile_model(model, inputs) -> 'ONNXModel':
oc.chainer_f_converter.clear()
oc.chainer_l_converter.clear()
oc.chainer_l_converter[L.Linear] = lb.convert_onnx_chainer_linear
oc.chainer_l_converter[L.Convolution2D] = lb.convert_onnx_chainer_convolution2d
oc.chainer_l_converter[L.BatchNormalization] = lb.convert_onnx_chainer_batch_normalization
oc.chainer_f_converter[F.relu] = fb.convert_relu
oc.chainer_f_converter[F.softmax] = fb.convert_softmax
oc.chainer_f_converter[F.pad_sequence] = fb.convert_pad_sequence
oc.chainer_f_converter[F.softmax_cross_entropy] = fb.convert_softmax_cross_entropy
oc.chainer_f_converter[F.average_pooling_2d] = fb.convert_average_pool_2d
oc.chainer_f_converter[F.unpooling_2d] = fb.convert_unpooling_2d
oc.chainer_f_converter[F.vstack] = fb.convert_vstack
oc.chainer_f_converter[F.hstack] = fb.convert_hstack
oc.chainer_f_converter[F.stack] = fb.convert_stack
oc.chainer_f_converter[F.separate] = fb.convert_separate
oc.chainer_f_converter[F.squeeze] = fb.convert_squeeze
oc.chainer_f_converter[F.reshape] = fb.convert_reshape
oc.chainer_f_converter[F.split_axis] = fb.convert_split_axis
oc.chainer_f_converter[F.swapaxes] = fb.convert_swapaxes
oc.chainer_f_converter[F.dropout] = fb.convert_dropout
oc.chainer_f_converter[F.matmul] = fb.convert_matmul
oc.chainer_f_converter[F.concat] = fb.convert_concat
oc.chainer_f_converter[F.max_pooling_2d] = fb.convert_max_pooling_2d
oc.chainer_f_converter[F.resize_images] = fb.convert_resize_images
oc.chainer_f_converter[F.tanh] = fb.convert_tanh
oc.chainer_f_converter[F.sigmoid] = fb.convert_sigmoid
oc.chainer_f_converter[F.broadcast_to] = fb.convert_broadcast_to
oc.chainer_f_converter[F.expand_dims] = fb.convert_expand_dims
oc.chainer_f_converter[F.local_response_normalization] = fb.convert_local_response_normalization
oc.chainer_f_converter[F.average] = fb.convert_average
if int(chainer.__version__[0]) >= 6:
oc.chainer_f_converter[F.roi_max_pooling_2d] = fb.convert_roi_max_pooling_2d
oc.chainer_f_converter[F.roi_average_pooling_2d] = fb.convert_roi_average_pooling_2d
oc.chainer_f_converter[F.roi_max_align_2d] = fb.convert_roi_max_align_2d
oc.chainer_f_converter[F.roi_average_align_2d] = fb.convert_roi_average_align_2d
# assign names
oc.assigned_names.clear()
oc.node2onnx_parameter.clear()
oc.value2onnx_parameter.clear()
inputs_, outputs_, graph_ = core.convert_model(model, inputs)
if graph_ is None:
return None
oc.preprocess(graph_, True)
generator = oc.ONNXGenerator()
model = generator.generate_model(
graph_.input_values, graph_.output_values, graph_, model)
# check inputs
onnx_model = ONNXModel()
onnx_model.model = model
onnx_model.inputs = graph_.input_values
onnx_model.outputs = graph_.output_values
return onnx_model
