现在我使用使用aclruntime推理om模型,程序如下“
data = np.array(data).astype(np.float32)
data = np.ascontiguousarray(data)
model_path="M1.om"
device_id = 0
options = aclruntime.session_options()
session = aclruntime.InferenceSession(model_path,device_id,options)
input_names = [input_meta.name for input_meta in session.get_inputs()]
output_names = [output_meta.name for output_meta in session.get_outputs()]
feeds = [aclruntime.Tensor(data)]
output = session.run(output_names,feeds) ”
打印output为“out1: <Tensor>
shape: (1, 20)
dtype: float32
device: 0
out2: <Tensor>
shape: (1, 1)
dtype: float32
device: 0
”我现在想知道out1和out2具体的数值,我改怎么做?
现在我使用使用aclruntime推理om模型,程序如下“
data = np.array(data).astype(np.float32)
data = np.ascontiguousarray(data)
model_path="M1.om"
device_id = 0
options = aclruntime.session_options()
session = aclruntime.InferenceSession(model_path,device_id,options)
input_names = [input_meta.name for input_meta in session.get_inputs()]
output_names = [output_meta.name for output_meta in session.get_outputs()]
feeds = [aclruntime.Tensor(data)]
output = session.run(output_names,feeds) ”
打印output为“out1: <Tensor>
shape: (1, 20)
dtype: float32
device: 0
out2: <Tensor>
shape: (1, 1)
dtype: float32
device: 0
”我现在想知道out1和out2具体的数值,我改怎么做?