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 2021/05/18 19:59:11 UTC

[GitHub] [tvm] manupa-arm edited a comment on issue #8062: AOT C Codegen Type Issue

manupa-arm edited a comment on issue #8062:
URL: https://github.com/apache/tvm/issues/8062#issuecomment-843513550


   @mehrdadh , can you confirm how do you know the output type is int8 ?
   Is this a quantized network ?
   
   I see symptoms of a different problem here because there is a fused cast at the last operator. For argument sake, can you usethe last operator's arg type (post-cast -- in _fused_divide_add_round_cast_clip_cast_) and see whether the issue persists. If my thinking is correct, the compiled model might expect the output type to be int_16 and not int_8.
   
   cc @giuseros : I think there might be int_16 cast at the very end causing this.


-- 
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