Exploring Posttraining Quantization of Large Language Models: An Efficiency Evaluation with a Focus on Russian-Language Tasks | Synapse