而您的模型中使用了int64类型的输入。您可以尝试将模型中的int64类型的输入转换为int32类型，然后重新...
转换出问题:Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32. Thanks for your bug report. We appreciate it a lot. Checklist I have searched related issues but cannot get the expected help. I have read th...
elementwise_floordiv 算子在 int32/int64 的情况下直接转换成了 ONNX 中的 div 算子,由于 div 算子是普通除操作,而不是整除操作,因此无法通过 CI 的校验。 2 实现过程 原核心实现代码如下 voidElementWiseFloordivMapper::Opset7(){autoinput_x_info =GetInput("X");autoinput_y_info =GetInput("Y");aut...
会提示 int32 和 int64 不兼容的问题 # token_num_int = torch.max(token_num).type(torch.int32).item() token_num_int = torch.max(token_num).type(torch.int64) acoustic_embeds = acoustic_embeds[:, :token_num_int, :]
to= TensorProto.INT64 ) new_nodes += [new_scale_node, new_add_node] else: new_nodes += [node] returnnew_nodes if__name__=='__main__': model = onnx.load('resize_conv_add.onnx') graph = model.graph nodes = graph.node ...
[] nrows =Noneifbatch_axiselse1fork, vinzip(df.columns, df.dtypes):ifdropisnotNoneandkindrop:continueifv =='int64': t = Int64TensorType([nrows,1])elifv =='float32': t = FloatTensorType([nrows,1])elifv =='float64': t = DoubleTensorType([nrows,1])else:raiseException("Bad type...
[] nrows =Noneifbatch_axiselse1fork, vinzip(df.columns, df.dtypes):ifdropisnotNoneandkindrop:continueifv =='int64': t = Int64TensorType([nrows,1])elifv =='float32': t = FloatTensorType([nrows,1])elifv =='float64': t = DoubleTensorType([nrows,1])else:raiseException("Bad type...
selected_indices:张量(int64)。从框张量中选择的索引 [num_selected_indices,3],选定的索引格式为[batch_index,class_index,box_index]。 我的检测器。2 个输出。框和分数 如你所见,这些确实是输出名称。你的模型可能有不同的名称,因此使用Netron是一种很好的做法。
tensor.type.tensor_type.shape.dim[0].dim_value=int(batch_size)fornodeingraph.node:ifnode.op_type !='Reshape':continueforinitingraph.initializer:ifinit.name != node.input[1]:continueiflen(init.int64_data) >0: init.int54_data[0]= -1eliflen(init.raw_data) >0: ...
Int8、Int16、Int32、Int64 Quantized Int uint8, uint16, uint32, uint64 Float16, float, double Bool String Complex64, complex128 非张量类型: Sequence Map Operators (Built-in/ Custom) 📌 算子/运算符 ONNX Graph 中的每个计算节点都执行特定的操作并产生一个或多个输出。 ONNX 标准定义了 📘...