[torchlib] Implement RNN operators

[justinchuby]

• lstm

- func: lstm.input(Tensor input, Tensor[] hx, Tensor[] params, bool has_biases, int num_layers, float dropout, bool train, bool bidirectional, bool batch_first) -> (Tensor, Tensor, Tensor)
  tags: nondeterministic_seeded

• gru

- func: gru.input(Tensor input, Tensor hx, Tensor[] params, bool has_biases, int num_layers, float dropout, bool train, bool bidirectional, bool batch_first) -> (Tensor, Tensor)
  tags: nondeterministic_seeded

Test models

• [export] nn.GRU fails to torch.export due to unimplemented operator pytorch/pytorch#120626
• [export] nn.LSTM fails to torch.export due to unimplemented operator pytorch/pytorch#123089

[utsab345]

I can work on this , Please assigned it to me.

[ombrdr47]

Hi @justinchuby! I'd like to work on implementing the LSTM and GRU operators for torchlib.

I have intermediate Python and PyTorch experience and I'm familiar with RNN architectures. I've reviewed the linked PyTorch issues and understand these operators are needed for ONNX export.

I plan to:

1. Study existing torchlib operator implementations
2. Implement lstm.input() and gru.input() with proper parameter handling
3. Add comprehensive tests

Could you confirm this is still available?

Thanks!