-
Notifications
You must be signed in to change notification settings - Fork 534
feat: Support training of RNN and LSTM. #1110
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Wanglongzhi2001
commented
Jun 18, 2023
•
edited
Loading
edited
- Support the model building of RNN.
- Support the model building of LSTM.
- Support the training of RNN.
- Support the training of LSTM.
- Add test.
- Refactor Tensors class to deal with implicit conversion from Tensors to Tensor and support Nest Tensors.
- Add Nest class to handle nested Tensor cases, and add corresponding Flatten and Mapstructure API.
- Use IOptionalArgs in Layer.Apply to prevent the inconsistency of Call method parameters in different layers, resulting in the need to set too many Call method parameters.
- Add the implemention of while_loop in graph mode.
feat: add rnn basic modules
fix: some possible errors of RNN.
Add feature(not completed):add SimpleRNNCell, StackedRNNCell, RNN and test.
feat: support training of RNN.
It builds on windows successfully, but fails to build on linux. According to the error message, it seems that the problem is on protobuf. Is this matters? @AsakusaRinne |
I'll test it on linux later |
The problem in ci has been located, which will be solved after upgrading the redist packages later. |