Add blog link

README.md

@@ -15,6 +15,8 @@ metrics:
 
 This is an optimized model using [madlag/bert-large-uncased-wwm-squadv2-x2.63-f82.6-d16-hybrid-v1](https://huggingface.co/madlag/bert-large-uncased-wwm-squadv2-x2.63-f82.6-d16-hybrid-v1) as the base model which was created using the [nn_pruning](https://github.com/huggingface/nn_pruning) python library. This is a pruned model of [madlag/bert-large-uncased-whole-word-masking-finetuned-squadv2](https://huggingface.co/madlag/bert-large-uncased-whole-word-masking-finetuned-squadv2)
 
+Feel free to read our blog about how we optimized this model [(link)](https://tryolabs.com/blog/2022/11/24/transformer-based-model-for-faster-inference)
+
 Our final optimized model weighs **579 MB**, has an inference speed of **18.184 ms** on a Tesla T4 and has a performance of **82.68%** best F1. Below there is a comparison for each base model:
 
 | Model  | Weight | Throughput on Tesla T4 | Best F1 |