Lil'Log ยท 2023-01-10

Large Transformer Model Inference Optimization

<p><span class=&#34;update&#34;>[Updated on 2023-01-24: add a small section on <a href=&#34;#distillation&#34;>Distillation</a>.]</span><br/></p> <p>Large transformer models are mainstream nowadays, creating SoTA results for a variety of tasks. They are powerful but very expensive to train and use. The extremely high inference cost, in both time and memory, is a big bottleneck for adopting a powerful transformer for solving real-world tasks at scale.</p> <p><strong>Why is it hard to run inference for large transformer models?</strong> Besides the increasing size of SoTA models, there are two main factors contributing to the inference challenge (<a href=&#34;https://arxiv.org/abs/2211.05102&#34;>Pope et al. 2022</a>):</p>

Open original