Noiz Ai

Thanks for giving a look! Let's ditch the audio stress together and make creating audio/video (or any content!

[Noiz Ai @noiz_ai]

Try the https://t.co/UTRf9uybPJ Hotel Lobby template here:

https://t.co/FQKjJrICki

Two photos → one short AI video.

Use your own photos, a friend, or an original character and make your version.

[Noiz Ai @noiz_ai]

you can even rewrite the lyrics in few simple steps!

[Noiz Ai @noiz_ai]

Try the https://t.co/UTRf9uybPJ Hotel Lobby template here:

https://t.co/FQKjJrICki

Two photos → one short AI video.

Use your own photos, a friend, or an original character and make your version.

[Aaliya @aaliya_va]

Two photos → a full AI video. That’s a wild use of a simple template.

@noiz_ai makes it easy to turn your own photos into short videos with this Hotel Lobby template.

Try it here:
https://t.co/tmh7nW4Z6V

#HotelLobby #aivideo

[Noiz Ai @noiz_ai]

Try the https://t.co/UTRf9uybPJ Hotel Lobby template here:

https://t.co/FQKjJrICki

Two photos → one short AI video.

Use your own photos, a friend, or an original character and make your version.

[Noiz Ai @noiz_ai]

try YuE2 in Noiz ai:
https://t.co/e7dP4pzH7S

[Noiz Ai @noiz_ai]

We’re proud to have contributed to YuE2 alongside HKUST, M-A-P and the broader research team.

What excites us most is music generation with an editable symbolic plan, not just a finished waveform. Open models are getting closer to real creative tools.

Congrats to the team on the release.

[AI Search @aisearchio]

New open source music generator, YuE2

Even beats Mureka v9 and Minimax 3 on various benchmarks

https://t.co/HD8q5DT0sG

[Ostris @ostrisai]

AI Toolkit now supports training LoRAs for YuE2-3B, an amazing music model. Support would not be possible without the tokenizer and help from @sin_ceriously . Thank you! Also special thanks to @Machinedelusion. More in 🧵

[TheAwakenOne @TheAwakenOne619]

I wanted to see how deep and emotional I could make a song with FL YuE2. My lil dude passed away about a month ago. I played this song and my daughter couldn't stop crying lol

I'm experimenting with connecting my DGX Spark running qwen3.8-flash-next to this workflow and using an image as a reference.

RIP Leo!
I miss my lil dude so much!

Now you can do Lora training. Need to give that a shot next!
https://t.co/L2y3rcxV68

[Adina Yakup @AdinaYakup]

YuE2 🎵 an open music model from the M-A-P community!

- Outperforming Suno v5/v6 in WildSongBench
- Edit melody & chords with ABC scores
- 3B, runs locally on a 24GB GPU
- Turn existing melodies into new song styles

[GitHubDaily @GitHub_Daily]

YuE2，港科大、纽约大学、斯坦福等几家联合出的开源音乐模型，给它歌词和风格描述，先写出一份旋律和和弦的谱，再按谱生成带人声伴奏的整首歌。

中间那份谱是能看能改的，改一段和声、换个速度、动几个音，再交回去生成新的一版，整首歌怎么走自己说了算。

GitHub：https://t.co/lm7zPgKD3X

翻唱也能做，把一段现成录音转成旋律谱，配新歌词或者换个风格，同一个模型直接出一版新的演绎。

官方演示里一首《The Last Train》通过对话改了 9 步 14 个版本，从中文流行一路改成英文爵士还加了段萨克斯独奏，每一版的谱和对话都能翻。

自家评测里跟 Suno v5、v6 打得有来有回，README 也承认头几名差距很小分不出高下。

配套给了一个 Agent Skill，让 Claude Code 这类工具直接调它写歌改歌。

需要 Linux 加 24 GB 显存的 N 卡，出的是 48 kHz 立体声。个人和音乐人自己用是免费的，生成的作品拿去变现也不用交授权费。

[TheAwakenOne @TheAwakenOne619]

I asked qwen3.8-flash-next to write a song about itself lol

Music: YuE2
Visualizer: PulseForge
Both via Pinokio

Compose:
English, gangsta rap over church organ, male preacher-trap vocal, 808s, choir hook, funeral and flex, 138 BPM

Generate Text:
Write YuE2 song lyrics only. No title, no subtitle, no markdown, no commentary, do not use words from the style prompt. Use exactly these section tags on their own lines. DO NOT USE STYLE PROMPT IN THE LYRICS. DO NOT USE FL YUE2 COMPOSE IN THE LYRICS: 2x [verse] [chorus] [bridge] [outro]. Two verses and two choruses. Rap about you and what you can do, you are qwen3.8-flash-next running in hermes, funny parody. English, gangsta rap over church organ

[Machine Delusions @Machinedelusion]

Training for YuE2 now in comfy with my custom node pack!

Shout out to @sin_ceriously for the RVQ Adapter that made this possible

https://t.co/4T2hC3qY6h

[Rebel AI @realrebelai]

Yue2 can completely restyle songs if you didnt know.

heres a sample of a fire example. notice the cadence, tempo, melodic background instruments and vocals.

Were finally here boys.

Khelani - When Hes Not There (Male Vocal Restyle)

[カイトラ (Kytra) 凯托拉 @sin_ceriously]

The "hum-to-song" module training is complete. I'm open-sourcing it within the hour. Follow to make sure you don't miss the update.
Special thanks to my mates at https://t.co/dR2ZWfpaeK for sponsoring the compute.
#YuE2 #opensource

[AI Search @aisearchio]

New open source music generator, YuE2

Even beats Mureka v9 and Minimax 3 on various benchmarks

https://t.co/HD8q5DT0sG

[Noiz Ai @noiz_ai]

@zikasu250468 We are very sorry for the poor experience. We have just fixed the issue, so please refresh and check again.🫣

[Noiz Ai @noiz_ai]

Say goodbye to endless copy-pasting 🎙️ With NoizAI-TTS, import scripts, identify multiple speakers, and generate voiceover-all in one place.
#NoizAI #AITTS #AIVoice #VideoCreation #CreatorTools #ContentCreation

[Noiz Ai @noiz_ai]

We’re proud to have contributed to YuE2 alongside HKUST, M-A-P and the broader research team.

What excites us most is music generation with an editable symbolic plan, not just a finished waveform. Open models are getting closer to real creative tools.

Congrats to the team on the release.

[Noiz Ai @noiz_ai]

@cie7le 氷の湖を4回渡ってロシア国境まで行くという具体性があって、写真だけより旅のスケールが伝わりますね。費用や日数まで書かれているので、実際に行くイメージを持ちやすいです。

[Noiz Ai @noiz_ai]

@jiro_3_ 「本当に辞めるの？」と何度も確認される流れがリアルですね。転職を決めた後に必要なのは気持ちの説得ではなく、判断を整理して次に進むことだと伝わります。

[Noiz Ai @noiz_ai]

@musitsune @dreamina_ai ベルサイユ庭園の雰囲気を20秒の散歩に落とし込んでいるのが素敵ですね。キャラクターが歩く時間と場所の空気感があるので、単なるAI画像の紹介よりもIPの世界に入りやすいです。