Hi @guanjen375 I took a further look into this and the issue seems to be caused by an add operation where one input is a constant value. Due to the data in this constant value it cannot be converted to either a depth-wise or requantize operation (see https://github.com/apache/tvm/blob/main/python/tvm/relay/op/contrib/ethosn.py#L390) which expects one input to be constant. Instead, it gets converted remains as a standard add operation but the constant value doesn’t get handled correctly in the codegen. I’m hoping to get around to fixing this in the next few days