As a mainstream approach in the quantum machine learning field, variational algorithms (VQAs) are frequently mentioned among most promising applications for computing. However, VQAs suffer from inefficient training methods. Here, we propose pretraining strategy named near Clifford circuits warm start (NCC-WS) to find initialization parameterized (PQCs) VQAs. We explored expressibility of NCCs a...