yongwww commented on PR #14123:
URL: https://github.com/apache/tvm/pull/14123#issuecomment-1446683807
> Thanks @yongwww , as a followup, would be great if we can also confirm if
it works on TVMScript parser
Thanks @tqchen! Is the following test case what you meant?
```python
@tvm.script.ir_module
class Before:
@R.function
def main(x: R.Tensor((10, 20), "float32")) -> R.Tensor(None,
"float32", ndim=2):
# block 0
with R.dataflow():
lv0 = R.emit_te(topi.add, x, relax.const(1, "float32"))
lv1 = R.emit_te(topi.exp, lv0)
gv = R.emit_te(topi.squeeze, lv1)
R.output(gv)
return gv
Before.show()
```
output of Before.show():
```python
@I.ir_module
class Module:
@R.function
def main(
x: R.Tensor((10, 20), dtype="float32")
) -> R.Tensor((10, 20), dtype="float32"):
with R.dataflow():
lv = R.call_tir(
add,
(x, R.const(1, "float32")),
out_sinfo=R.Tensor((10, 20), dtype="float32"),
)
lv0: R.Tensor((10, 20), dtype="float32") = lv
lv2 = R.call_tir(exp, (lv0,), out_sinfo=R.Tensor((10, 20),
dtype="float32"))
lv1: R.Tensor((10, 20), dtype="float32") = lv2
lv4 = R.call_tir(
squeeze, (lv1,), out_sinfo=R.Tensor((10, 20),
dtype="float32")
)
gv: R.Tensor((10, 20), dtype="float32") = lv4
R.output(gv)
return gv
```
--
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]
