feat: add echo experiment

Author: flavioschneider

README.md

@@ -164,7 +164,8 @@ y_long = composer(y, keep_start=True) # [1, 1, 98304]
 | [Alpha](https://wandb.ai/schneider/audio/reports/Audio-Diffusion-UNet-Alpha---VmlldzoyMjk3MzIz?accessToken=y0l3igdvnm4ogn4d3ph3b0i8twwcf7meufbviwt15f0qtasyn1i14hg340bkk1te) | [6bd9279f19](https://github.com/archinetai/audio-diffusion-pytorch/tree/6bd9279f192fc0c11eb8a21cd919d9c41181bf35) | Initial tests on LJSpeech dataset with new architecture and basic DDPM diffusion model. |
 | [Bravo](https://wandb.ai/schneider/audio/reports/Audio-Diffusion-Bravo---VmlldzoyMzE4NjIx?accessToken=qt2w1jeqch9l5v3ffjns99p69jsmexk849dszyiennfbivgg396378u6ken2fm2d) | [a05f30aa94](https://github.com/archinetai/audio-diffusion-pytorch/tree/a05f30aa94e07600038d36cfb96f8492ef735a99) | Elucidated diffusion, improved architecture with patching, longer duration, initial good (unsupervised) results on LJSpeech.
 | [Charlie](https://wandb.ai/schneider/audio/reports/Audio-Diffusion-Charlie---VmlldzoyMzYyNDA1?accessToken=71gmurcwndv5e2abqrjnlh3n74j5555j3tycpd7h40tnv8fvb17k5pjkb57j9xxa) | [50ecc30d70](https://github.com/archinetai/audio-diffusion-pytorch/tree/50ecc30d70a211b92cb9c38d4b0250d7cc30533f) | Train on music with [YoutubeDataset](https://github.com/archinetai/audio-data-pytorch), larger patch tests for longer tracks, inpainting tests, initial test with infinite generation using SpanBySpanComposer. |
-| [Delta](https://wandb.ai/schneider/audio/reports/Audio-Diffusion-Delta---VmlldzoyNDYyMzk1?accessToken=n1d34n35qserpx7nhskkfdm1q12hlcxx1qcmfw5ypz53kjkzoh0ge2uvhshiseqx) | (current) | Test model with the faster `ADPM2` sampler and dynamic thresholding. |
+| [Delta](https://wandb.ai/schneider/audio/reports/Audio-Diffusion-Delta---VmlldzoyNDYyMzk1?accessToken=n1d34n35qserpx7nhskkfdm1q12hlcxx1qcmfw5ypz53kjkzoh0ge2uvhshiseqx) | [672876bf13](https://github.com/archinetai/audio-diffusion-pytorch/tree/672876bf1373b1f10afd3adc8b3b984495bca91a) | Test model with the faster `ADPM2` sampler and dynamic thresholding. |
+| [Echo](https://wandb.ai/schneider/audio/reports/Audio-Diffusion-Echo---VmlldzoyNTU2NTcw?accessToken=sthdn25n8is30gjo2x0w4fs9hwbua23rlbg7o4bv8h17y47xdtruiiyb33aoc5h4) | (current) | Test `AudioDiffusionUpsampler`.
 
 ## TODO