[错误反馈] 04课Note代码 ValueError: 'relax.alloc_tensor' is not a valid SpecialOp

[Aguin]

在运行04课以下代码时

@R.function
def main(x: Tensor((1, 784), "float32"),
         w0: Tensor((128, 784), "float32"),
         b0: Tensor((128,), "float32"),
         w1: Tensor((10, 128), "float32"),
         b1: Tensor((10,), "float32")):

    with R.dataflow():
        lv0 = R.call_tir(linear0, (x, w0, b0), (1, 128), dtype="float32")
        gv0 = R.call_tir(relu0, (lv0,), (1, 128), dtype="float32")
        R.output(gv0)

    gv1 = R.alloc_tensor((1, 128), dtype="float32")

    with R.dataflow():
        out = R.call_tir(linear1, (gv0, gv1, b0), (1, 128), dtype="float32")
        R.output(out)
    return out

出现错误

ValueError                                Traceback (most recent call last)
ValueError: 'relax.alloc_tensor' is not a valid SpecialOp

During handling of the above exception, another exception occurred:

ValueError                                Traceback (most recent call last)
/data/venv/tvm/lib/python3.7/site-packages/tvm/script/relax/parser.py in parse_attr(self, expr)
   1033             try:
-> 1034                 return SpecialOp(op_name)
   1035             except ValueError:

/usr/lib/python3.7/enum.py in __call__(cls, value, names, module, qualname, type, start)
    314         if names is None:  # simple value lookup
--> 315             return cls.__new__(cls, value)
    316         # otherwise, functional API: we're creating a new Enum type

/usr/lib/python3.7/enum.py in __new__(cls, value)
    568             exc.__context__ = ve_exc
--> 569             raise exc
    570 

/usr/lib/python3.7/enum.py in __new__(cls, value)
    552             exc = None
--> 553             result = cls._missing_(value)
...

使用的环境是WSL(Ubuntu18)+Python3.7, 在Jupyter运行, mlc-ai-nightly-cu102版本0.9.dev1664+g1f3985de0

目前的解决方法是把R.alloc_tensor更改为R.builtin.alloc_tensor

想请问下出现这个问题的原因, 谢谢助教

[tqchen]

感谢指出问题。这个是我们的疏忽