Lyken17 edited a comment on pull request #10439:
URL: https://github.com/apache/tvm/pull/10439#issuecomment-1056656951
Also when I try to build model with relay API, such dtype still cannot be
handled properly.
```python
input = relay.var("input", shape=[1,3,32,32], dtype="float32")
weight = relay.var("weight", shape=[3,1,3,3], dtype="float32")
out = relay.nn.conv2d(input, weight, groups=3, channels=3)
fn = relay.Function([input, weight], out)
fmod = tvm.IRModule.from_expr(fn)
mod = relay.transform.InferType()(fmod)
bwd_expr = relay.transform.gradient(mod["main"], mode="first_order")
bwd_mod = tvm.IRModule.from_expr(bwd_expr)
bwd_mod = relay.transform.InferType()(bwd_mod)
```
```bash
data types float32 and void do not match in BroadcastRel
note: run with `TVM_BACKTRACE=1` environment variable to display a backtrace.
```
There might be some missing part that make `out_type` null in this case. For
now, I would recommend put "float32" as a default value to avoid these errors.
--
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]
