I have worked as a postdoctoral fellow in China for two years. My experience was mainly with two institutions -- CAS/Beijing and SCAU/Guangzhou --, but I have visited a few others informally. Thus I write here from my experience as a visiting scholar.
My impression is that the academia in China is strongly pressed to produce papers, and that all sorts of strategies are being employed by opportunistic PIs in face of the prospect of large short-term gains and fast career ascension. Most labs I have visited were crowded, and working 12/7, managed by 1-2 PIs only intermittently present. There is a growing interest in hiring external postdocs, where institutions are competing in salary conditions (often not met).
At the CAS I worked in a lab with many students, each working on different projects in diverse topics. Whilst the local PI had a strong physical presence and held constant meetings, all students complained they were being demanded for results without any actual support nor apprenticeship. The numerous students were also demanded to manage the PI's paperwork, post deliveries (most not work-related), peer-reviews, and paper submission steps. Often on Sundays. At one period that lab had 18 students sharing the same 2-room apartment, payed from their salary... to the PI.
I heard such lab situations are quite common currently, and that fitted with my witnessing of crowded labs. I do not think this local practice is academically nor scientifically healthy.
Therefore my main advice is that you search for a more balanced laboratory, perhaps considering a different culture / location towards a more solid, sane formation. If you find yourself in a bad place, don't play their game and leave it at once.
I just realised I did not answer your question objectively. The best labs I have dealt with had ca. 3-5 students per professor, most of them are typically Master students. Postdocs are independent and may assist in tutoring students, writing papers; I have seen well-reputed labs with up to 4 postdocs under the same supervisor.