You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by "arangasa (via GitHub)" <gi...@apache.org> on 2023/07/07 05:31:29 UTC
[GitHub] [tvm] arangasa opened a new pull request, #15258: [Hexagon] Propagate qnn.concat quantization params to inputs, elimina…
arangasa opened a new pull request, #15258:
URL: https://github.com/apache/tvm/pull/15258
…ting redundant requantization when possible, and make it concat
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] arangasa commented on pull request #15258: [Hexagon] Propagate qnn.concat quantization params to inputs, elimina…
Posted by "arangasa (via GitHub)" <gi...@apache.org>.
arangasa commented on PR #15258:
URL: https://github.com/apache/tvm/pull/15258#issuecomment-1624771971
cc @kparzysz-quic @quic-sanirudh
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] tvm-bot commented on pull request #15258: [Hexagon] Propagate qnn.concat quantization params to inputs, elimina…
Posted by "tvm-bot (via GitHub)" <gi...@apache.org>.
tvm-bot commented on PR #15258:
URL: https://github.com/apache/tvm/pull/15258#issuecomment-1624771312
<!---bot-comment-->
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from [Reviewers](https://github.com/apache/incubator-tvm/blob/master/CONTRIBUTORS.md#reviewers) by @-ing them in a comment.
<!--bot-comment-ccs-start-->
* cc @ibsidorenko <sub>See [#10317](https://github.com/apache/tvm/issues/10317) for details</sub><!--bot-comment-ccs-end-->
<sub>Generated by [tvm-bot](https://github.com/apache/tvm/blob/main/ci/README.md#github-actions)</sub>
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] quic-sanirudh commented on a diff in pull request #15258: [Hexagon] Propagate QNN Concat Quantization Params to Inputs
Posted by "quic-sanirudh (via GitHub)" <gi...@apache.org>.
quic-sanirudh commented on code in PR #15258:
URL: https://github.com/apache/tvm/pull/15258#discussion_r1255629596
##########
python/tvm/contrib/hexagon/transform.py:
##########
@@ -311,3 +318,73 @@ def remove_empty_pad(mod):
"""Remove the empty pad operator."""
mod["main"] = rewrite(remove_empty_pad_callback(), mod["main"])
return mod
+
+
+class simplify_qnn_concat_in_func(DFPatternCallback):
+
+ """
+ Propagate qnn.concat's quantization params to its inputs,
+ and try to avoid redundant requantization while doing so.
+ """
Review Comment:
Minor comment. If possible we could add a snippet of the relay IR before and after this pass in the comments similar to the comments mentioned in the functions above (like `remove_empty_pad_callback` or `qdistilbert_rewrite`).
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] quic-sanirudh merged pull request #15258: [Hexagon] Propagate QNN Concat Quantization Params to Inputs
Posted by "quic-sanirudh (via GitHub)" <gi...@apache.org>.
quic-sanirudh merged PR #15258:
URL: https://github.com/apache/tvm/pull/15258
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] quic-sanirudh commented on a diff in pull request #15258: [Hexagon] Propagate QNN Concat Quantization Params to Inputs
Posted by "quic-sanirudh (via GitHub)" <gi...@apache.org>.
quic-sanirudh commented on code in PR #15258:
URL: https://github.com/apache/tvm/pull/15258#discussion_r1255629596
##########
python/tvm/contrib/hexagon/transform.py:
##########
@@ -311,3 +318,73 @@ def remove_empty_pad(mod):
"""Remove the empty pad operator."""
mod["main"] = rewrite(remove_empty_pad_callback(), mod["main"])
return mod
+
+
+class simplify_qnn_concat_in_func(DFPatternCallback):
+
+ """
+ Propagate qnn.concat's quantization params to its inputs,
+ and try to avoid redundant requantization while doing so.
+ """
Review Comment:
Minor comment. If possible could you add a snippet of the relay IR before and after this pass in the comments similar to the comments mentioned in the functions above (like `remove_empty_pad_callback` or `qdistilbert_rewrite`).
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [tvm] quic-sanirudh commented on pull request #15258: [Hexagon] Propagate QNN Concat Quantization Params to Inputs
Posted by "quic-sanirudh (via GitHub)" <gi...@apache.org>.
quic-sanirudh commented on PR #15258:
URL: https://github.com/apache/tvm/pull/15258#issuecomment-1628359815
@tvm-bot rerun
--
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: commits-unsubscribe@tvm.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org