abhikran-quic opened a new issue, #13133:
URL: https://github.com/apache/tvm/issues/13133
### Expected behavior
When using `relay.Constant` in `tvm.parser.fromtext` , `IRModule` should be
returned.
### Actual behavior
When using `relay.Constant` in `tvm.parser.fromtext` , the following error
is seen.
`no entry in the meta table for relay.Constant `
### Environment
Version: 0.10.0
Python-3.8.5
### Steps to reproduce
```
import tvm
import tvm.contrib.hexagon
from tvm import relay
mod = tvm.parser.fromtext(
"""
#[version = "0.0.5"]
def @main(%p0239: Tensor[(1, 12, 128, 128), float32], %input_40: Tensor[(1,
224, 224, 3), float32] /* ty=Tensor[(1, 224, 224, 3), float32] */) {
%0 = cast(%input_40, dtype="float16") /* ty=Tensor[(1, 224, 224, 3),
float16] */;
%1 = transpose(%0, axes=[0, 3, 1, 2]) /* ty=Tensor[(1, 3, 224, 224),
float16] */;
%2 = layout_transform(%1, src_layout="NCHW", dst_layout="NHWC") /*
ty=Tensor[(1, 224, 224, 3), float16] */;
%3 = layout_transform(meta[relay.Constant][0] /* ty=Tensor[(64, 3, 3, 3),
float16] */, src_layout="OIHW", dst_layout="HWIO") /* ty=Tensor[(3, 3, 3, 64),
float16] */;
%4 = expand_dims(meta[relay.Constant][1] /* ty=Tensor[(64), float16] */,
axis=1, num_newaxis=2) /* ty=Tensor[(64, 1, 1), float16] */;
cast(%p0239, dtype="int8") /* ty=Tensor[(1, 12, 128, 128), int8] */
}
""")
params = {}
target_hexagon = tvm.target.hexagon("v68")
with tvm.transform.PassContext(opt_level=3):
hexagon_lowered = relay.build(
mod,
tvm.target.Target(target_hexagon, host=target_hexagon),
params=params,
)
```
### Triage
* needs-triage
* relay:parser
--
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]
