FrozenGene commented on a change in pull request #4828: [QNN][TFLite] TFLite
rounding mode support
URL: https://github.com/apache/incubator-tvm/pull/4828#discussion_r376237528
##########
File path: python/tvm/relay/frontend/tflite.py
##########
@@ -1212,6 +1214,7 @@ def convert_conv(self, op, conv_type):
input_zero_point=new_input_zero_point,
output_scale=output_tensor.qnn_params['scale'],
output_zero_point=output_tensor.qnn_params['zero_point'],
+ rounding="TFLITE",
Review comment:
As the comment said, we could make the `rounding` is an argument of
`from_tflite`, the default value is `TFLITE`. i.e.
```python
def from_tflite(model, shape_dict, dtype_dict, rounding="TFLITE"):
```
Of course, you should pass the `rounding` to `OperatorConverter` as one of
its member value so that other `convert_xxx` functions could get the value
----------------------------------------------------------------
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]
With regards,
Apache Git Services
