Integer Only Quantization and Pruning of Transformers