Update image

Author: curiousily

manuscript/06.time-series-anomaly-ecg.md

@@ -189,7 +189,7 @@ test_anomaly_dataset, _, _ = create_dataset(anomaly_df)
 
 ### LSTM Autoencoder
 
-![Autoencoder](https://lilianweng.github.io/lil-log/assets/images/autoencoder-architecture.png)
+![png](images/pytorch-05/autoencoder-architecture.png)
 _Sample Autoencoder Architecture [Image Source](https://lilianweng.github.io/lil-log/2018/08/12/from-autoencoder-to-beta-vae.html)_
 
 The general Autoencoder architecture consists of two components. An _Encoder_ that compresses the input and a _Decoder_ that tries to reconstruct it.