MasterJH5574 commented on pull request #10789:
URL: https://github.com/apache/tvm/pull/10789#issuecomment-1079625981
Hi @shingjan, I noted that our TVMScript parser fails when `T.int64` appears
in the parameter type annotation.
For example, when parsing the following script,
```python
@T.prim_func
def elementwise_shape_int64(
A: T.Buffer[(T.int64(128), T.int64(128)), "float32"],
C: T.Buffer[(T.int64(128), T.int64(128)), "float32"],
) -> None:
B = T.alloc_buffer((T.int64(128), T.int64(128)), dtype="float32")
for i, j in T.grid(128, 128):
with T.block("B"):
vi, vj = T.axis.remap("SS", [i, j])
B[vi, vj] = A[vi, vj] * 2.0
for i, j in T.grid(T.int64(128), T.int64(128)):
with T.block("C"):
vi, vj = T.axis.remap("SS", [i, j])
C[vi, vj] = B[vi, vj] + 1.0
```
the parser fails with message
```
Traceback (most recent call last):
File "tensorir.py", line 32, in Fused
A: T.Buffer[(T.int64(128), T.int64(128)), "float32"],
TypeError: __call__() takes 1 positional argument but 2 were given
```
Could you help take a look into this case? If the parser cannot support
`int64` in type annotation, we shouldn't print the buffer in sugar-style in the
printer. Thanks a lot!!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
