TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

MusiCoT, a chain-of-thought (CoT) prompting technique for music generation [pdf]

51 pointsby jinqueenyabout 1 month ago

6 comments

heeen2about 1 month ago
I can't comment on the merit of the technical aspects, but I feel like of all the AI generated content, especially AI generated music is as interesting as AI generated memoirs - sort of pointless. It lacks the human element that makes it relatable on an emotional level.
评论 #43643322 未加载
评论 #43642949 未加载
评论 #43642935 未加载
评论 #43642421 未加载
评论 #43642787 未加载
MSFT_Edgingabout 1 month ago
So when is AI going to do useful work no one wants to do instead of doing the art people enjoy doing.
评论 #43643926 未加载
评论 #43643536 未加载
评论 #43643666 未加载
评论 #43643089 未加载
评论 #43643228 未加载
gnabgibabout 1 month ago
Do you think you&#x27;re, perhaps over doing self-the promotion?<p>&gt; Please don&#x27;t use HN primarily for promotion. It&#x27;s ok to post your own stuff part of the time, but the primary use of the site should be for curiosity.<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;newsguidelines.html">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;newsguidelines.html</a><p>(4 subs, 2 weeks) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=musicot.github.io">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=musicot.github.io</a><p>(3 subs, 1 week&#x2F;5 subs, 1 month) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=github.com%2Finclusionai">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=github.com%2Finclusio...</a><p>(3 subs 1 week) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=mainfunc.ai">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=mainfunc.ai</a><p>(2 subs, 2 weeks) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=mureka.ai">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=mureka.ai</a><p>(4 subs, 3 months) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=trae.ai">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=trae.ai</a><p>(approaching ∞ subs) <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=pingcap.com">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;from?site=pingcap.com</a>
评论 #43642588 未加载
评论 #43643368 未加载
评论 #43641714 未加载
评论 #43641624 未加载
评论 #43642963 未加载
评论 #43642663 未加载
zaptremabout 1 month ago
I work on music models, and this is a very cool paper! There are no papers that go into depth on how token-based AR music models (that aren&#x27;t absurdly inefficient like Yue) are trained. I&#x27;m particularly interested in your semantic tokens. I tried reproducing the CTC loss part but my curve was very spikey and didn&#x27;t seem to actually figure out any characters. The semantic tokens gave great acoustic info but gibberish lyrics. What did your CTC loss curves look like and did you see anything similar at any point?<p>As a semi-aside, I feel like semantic tokens in general may end up being a bottleneck on how interesting model outputs can be.
评论 #43643518 未加载
评论 #43643878 未加载
AIPedantabout 1 month ago
Were there any examples of &quot;de novo&quot; music generation using this? The only one I could find on the website was translating the vocals of an existing song, couldn&#x27;t find any AI compositions.
yxchngabout 1 month ago
This company has a track record of cheating on benchmark by training on test datasets. Take it with pinch of salt and try the model yourself.
评论 #43642128 未加载
评论 #43642995 未加载