From 6a314ea7cd01a78a58403bc83e7c637ef83e6b26 Mon Sep 17 00:00:00 2001
From: Blake Wyatt <894305+xNul@users.noreply.github.com>
Date: Tue, 22 Aug 2023 02:22:45 -0400
Subject: [PATCH] [DOCS] MusicGen Docs Update (#25510)

* docs: note token limitations for MusicGen

* docs: note token limitations for MusicGen

* docs: fix token count with token limitations for MusicGen
---
 docs/source/en/model_doc/musicgen.md | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/docs/source/en/model_doc/musicgen.md b/docs/source/en/model_doc/musicgen.md
index 72250a86fc..f93466c629 100644
--- a/docs/source/en/model_doc/musicgen.md
+++ b/docs/source/en/model_doc/musicgen.md
@@ -53,6 +53,10 @@ better results than greedy, thus we encourage sampling mode to be used where pos
 and can be explicitly specified by setting `do_sample=True` in the call to [`MusicgenForConditionalGeneration.generate`],
 or by overriding the model's generation config (see below).
 
+Generation is limited by the sinusoidal positional embeddings to 30 second inputs. Meaning, MusicGen cannot generate more
+than 30 seconds of audio (1503 tokens), and input audio passed by Audio-Prompted Generation contributes to this limit so,
+given an input of 20 seconds of audio, MusicGen cannot generate more than 10 seconds of additional audio.
+
 ### Unconditional Generation
 
 The inputs for unconditional (or 'null') generation can be obtained through the method