vicalloy opened a new issue #7886:
URL: https://github.com/apache/tvm/issues/7886
1. I get the follow error when run `relay.build`.
```
Check failed: (pval != nullptr) is false: Cannot allocate memory symbolic
tensor shape [?, ?]
```
2. The follow code is used to build the model.
```python
# model from openvino.
https://github.com/openvinotoolkit/open_model_zoo/blob/master/models/intel/face-detection-0200/description/face-detection-0200.md
onnx_model = onnx.load('face-detection-0200.onnx')
shape = [1,3,256,256]
input_name = "image"
shape_dict = {
input_name: shape,
}
mod, params = relay.frontend.from_onnx(onnx_model, shape_dict,
freeze_params=True)
print(relay.transform.DynamicToStatic()(mod))
with relay.build_config(opt_level=3):
graph, lib, params = relay.build(mod, target, params=params)
```
3. `dyn.strided_slice` output a dynamic shape.
```c
%329 = (%328, meta[relay.Constant][134] /* ty=Tensor[(1), int64] */);
%330 = concatenate(%329) /* ty=Tensor[(2), int64] */;
%331 = dyn.strided_slice(%327, meta[relay.Constant][133] /* ty=Tensor[(2),
int32] */, %330, meta[relay.Constant][135] /* ty=Tensor[(2), int32] */,
begin=None, end=None, strides=None) /* ty=Tensor[(?, ?), int64] */;
%332 = take(%331, meta[relay.Constant][136] /* ty=Tensor[(1), int64] */,
axis=1) /* ty=Tensor[(?, 1), int64] */;
%333 = squeeze(%332, axis=[1]) /* ty=Tensor[(?), int64] */;
%334 = cast(%333, dtype="int64") /* ty=Tensor[(?), int64] */;
%335 = cast(%334, dtype="int64") /* ty=Tensor[(?), int64] */;
%336 = cast(%335, dtype="int64") /* ty=Tensor[(?), int64] */;
```
--
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.
For queries about this service, please contact Infrastructure at:
[email protected]
