Reporting Strange behvaior (Error) when analyzing LSTM model with stm32ai (batch size is not 1) Hello,When I export the torch.nn.LSTM to ONNX and use onnxsim to simplify(and also eliminate onnx expand layer, which is not supprted) it,the generated .onnx model shows the following error after analyzing with stm32ai.exe (version 8.1.0)NOT IMPLEMENTED: Batch size greater than 1 not implementedI suspect that the transpose layer just before the LSTM misleads the stm32ai's analyzer to see 7 is batch size in the following onnx graph, which is the length of input sequence.And I wonder if I'm confused or missed the right way to convert the lstm model to onnx.For reproduction, I submit the code used for creating the model and lstm onnx file.the version of onnxsim is 0.4.33, pytorch is 2.0.0 import torch
import torch.nn as nn
class LSTMModel(nn.Module):
def __init__(self, vocab_size, embedding_dim, hidden_dim, output_dim, num_layers):
super(LSTMModel, self).__init__()
self.lstm = nn.LSTM(embedding_dim, hidden_dim, num_layers=num_layers, batch_first=True)
self.fc = nn.Linear(hidden_
