kice commented on issue #3879:
URL: https://github.com/apache/incubator-tvm/issues/3879#issuecomment-626418827
If you take a look here
```
auto inter_type = op->type.is_int() ? Int(16) : UInt(16);
value_int8 << CastFromTo(value_16.str(), op->value.type(), inter_type);
os << CastFromTo(value_int8.str(), inter_type, op->type);
```
We need to convert to `uint16` and then cast to `uint8`. No direct
conversion from `fp16` to `uint8`.
ref:
https://docs.nvidia.com/cuda/cuda-math-api/group__CUDA__MATH____HALF__MISC.html#group__CUDA__MATH____HALF__MISC
----------------------------------------------------------------
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]
