You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2020/02/10 10:24:42 UTC
[GitHub] [incubator-tvm] inadob edited a comment on issue #4805:
[Frontend][TFlite] Add parser support for relu6, leaky_relu, relu_n1_to_1,
log_softmax
inadob edited a comment on issue #4805: [Frontend][TFlite] Add parser support for relu6, leaky_relu, relu_n1_to_1, log_softmax
URL: https://github.com/apache/incubator-tvm/pull/4805#issuecomment-584053528
> Relu and Clip implementation does not look right.
>
> We can keep the computation in integer domain. The way to do that is to subtract the input zero point, and then call Relu, then requantize to the output scale (only if output scale/zero point are different from input scale/zero point).
Is that lowering to int8, not int32? @anijain2305
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services