This paper deals with the consensus problem in network of agents with cooperative and antagonistic interactions subject to quantization. By employing the techniques from nonsmooth analysis, we prove that all agents can be guaranteed to asymptotically reach bipartite consensus for any logarithmic quantizer accuracy under connected and structurally balanced topology and the states of all agents asymptotically converge to zero under connected and structurally unbalanced topology. In addition, finite-time bipartite consensus is considered for single-integrator agents with binary quantized information. The simulation results are given to demonstrate the effectiveness of the theoretical results.