fix(torch-frontend): Fixed torch.lstm to get correct attention output…

…s in the packed case.

ivy/functional/frontends/torch/nn/functional/layer_functions.py

@@ -107,6 +107,9 @@ def _generic_lstm(
     h_outs = h_out if num_layers == 1 else ivy.concat(h_outs, axis=0)
     c_outs = c_out if num_layers == 1 else ivy.concat(c_outs, axis=0)
 
+    if batch_sizes is not None:
+        output = _pack_padded_sequence(output, batch_sizes)[0]
+
     return output, h_outs, c_outs
 
 
@@ -160,10 +163,9 @@ def _lstm_cell(
         output = ivy.concat(ht_list, axis=0)
     else:
         ct_list = ivy.concat(ct_list, axis=0)
-        ht_list = ivy.concat(ht_list, axis=0)
+        output = ht_list = ivy.concat(ht_list, axis=0)
         c = _extract_states(ct_list, batch_sizes)
         h = _extract_states(ht_list, batch_sizes)
-        output = _pack_padded_sequence(ht_list, batch_sizes)[0]
     return output, (h, c)