Skip to content

Conversation

@zhang-yi-chi
Copy link
Contributor

@zhang-yi-chi zhang-yi-chi commented Jul 28, 2022

Add RNN operator for ONNX frontend. Note that currently onnxruntime (<= 1.12.0) has wrong default values (all zeros) for all activations so I removed default value tests for HardSigmoid.

cc @mbrookhart @AndrewZhaoLuo

@AndrewZhaoLuo
Copy link
Contributor

I'll take a look today

Copy link
Contributor

@AndrewZhaoLuo AndrewZhaoLuo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it looks fine. My main concern is about the high amount of similar code between this and LSTM. Can you try a refactor to share relevant parts with each other instead of copy paste?

Also q about tests

Copy link
Contributor

@AndrewZhaoLuo AndrewZhaoLuo left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thanks!

@AndrewZhaoLuo AndrewZhaoLuo merged commit cfa5525 into apache:main Aug 4, 2022
xinetzone pushed a commit to daobook/tvm that referenced this pull request Nov 25, 2022
…2213)

* Add RNN operation for ONNX frontend.

* link checks

* rm test_rnn_batchwise in unsupported_onnx_tests

* merge similar codes to class methods

* implement opset 14 and refactor test_forward

* reformat verify_rnn_helper

Co-authored-by: 张亦驰 <zhangyichi1@corp.netease.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants