In case anyone's wondering, performance testing results for running bf16 variant of #YuE2 music generation model (the open-source #suno alternative) in #ComfyUI on #RTX5090 - and yes, I'm running on the previous generation PC with DDR4 RAM, etc... but still...

ALT ComfyUI benchmark results for the YuE2 music generation model using an NVIDIA GeForce RTX 5090 GPU. After 1 warmup run, 10 benchmark runs completed successfully with no failures. Fastest generation time: 14.342 seconds; slowest: 21.823 seconds; average: 17.515 seconds; median: 17.190 seconds. The benchmark used the audio_yue2_text2music.json workflow. System specifications: NVIDIA RTX 5090 with 31.8 GB VRAM, AMD Ryzen 9 5950X 16-core processor (32 logical cores), 63.9 GB RAM, PyTorch 2.10.0+cu130, and Windows 11 Pro. A horizontal bar chart compares the fastest, slowest, average, and median execution times.
An endless loop of #DonaldTrump'sterfire – at Maardu, Eesti
When you accidentally load the CLIP (text encoder) from a different video model in ComfyUI, English turns into gibberish.
Since I already logged on, I might as well leave this here.
github.com/harlyh/stereo-pho…
The fully open-source music-generation model, #YuE2, is just amazing. You only need ~6GB of #GPU VRAM to run, and on beefier computers it's faster than #SUNO :)
Here are my #ComfyUI benchmark results. The test generated 1 (for warm-up) + 5 random songs, ranging from 1:38 to 2:00 in length.
On my PC, it generates at 10x the real-time speed.
Yesterday morning, I woke up with no idea that I'd end up upgrading my GPU to an RTX 9050 that same day.
Only after the upgrade did I start thinking about what I would use it for. As I couldn't figure it out yet, I used it to create this:
youtube.com/watch?v=DTCR6TZo…
Harly retweeted
TALES OF GRIMWICKE — Chapter I is 22 minutes of AI fantasy animation inspired by the many anthropomorphic animal stories I read as a kid, with a dark twist:
The model: Seedance 2.5 (@ByteDanceSeed_ ) via @Segmind_ai. 847 takes generated, mostly 30-second multi-shot prompts, over 5 hours of footage. 127 clips made the cut. About $4.5k in generation, four weeks from first take to publish.
Characters and locations came out of @midjourney. Using MJ for the location plates was a revelation: the cinematography and art direction are a level above, and it renders far better as a plate than Nano Banana or GPT Image. Nano Banana (@GoogleDeepMind run on @fal) did the heavy lifting editing those plates, building anchor frames and turning MJ characters into turnaround sheets. Then I let Seedance do as much of the framing and camera operating as possible. That gave the best results.
A lot of what I do in my Unreal Engine films came with me: a handheld, imperfect camera operator instead of smooth moves, focus hunting, and above all weather and atmospherics. The model does water, snow, smoke and rain really well, and leaning into that is what keeps it from looking like a Temu Pixar film.
The tools I built: a portal (Node on @flydotio) that holds the bible, the refs, every prompt and every take, with a Claude agent (@claudeai) that writes the prompts to the house rules. A Premiere Pro (@AdobeVideo) panel that browses the archive, drops selects on the timeline, versions the project, strips music, and sends fixes back to Seedance. Everything is done in 720p with no upscaling, because it looks terrible. Narration is my own voice through @ElevenLabs speech-to-speech. Music from @Artlist_io.
None of this would exist without @PJaccetturo and @Kavanthekid. Their narrative work, Nexus and Chronicles of Bone, proved long-form AI film was possible, and their podcasts, newsletters and X posts, updated constantly as the workflow changes under everyone's feet, showed how to actually do it.
Chapter II is currently in the works and I will release it initially on my YouTube channel, The Darkest Age (@CharlieDriscollAI). Consider subscribing there to see it as soon as it drops!
Apparently atmospheric disturbances combined with a shitty camera and lack of photography skills combined = angels 😆
As time goes by, I'm becoming even more convinced that the Bible is a poison for your mind. One step away from becoming completely braindead. – at Maardu, Eesti
Okay, @suno, I'm actually impressed. I was playing around with #suno #studio and wanted to apply and automate an #audio effect on a track. But my desired effect plugin was missing from the plugin selection.
What's impressive is that it wasn't a dealbreaker.
I could tell Studio's AI to code up a new plugin for me, exactly to my specification.
After 3 minutes of "DSP Sandbox" work, the AI came back with the freshly created plugin. And it freaking works. So, dragged it onto the track, and does exactly what I wanted.
Nice one! ;)
So let me get this straight... A dome somewhere out there is causing the Rainbow to curve even when you create a rainbow indoors, which is easy to do and has been demonstrated ad nauseam.
So exactly what is the mechanism? Oh, sorry, I forgot... You can't really answer the question because you dropped out of school just before the other kids started learning science... – at Maardu, Eesti
This Post is from a suspended account. Learn more
I'm not a veterinarian... but I think this guy over here needs to go back on his medications, or perhaps he needs to up the dose if he's currently on some already...
#flatearth #unhinged – at Maardu, Eesti
Replying to @sly_sparkane
Well, fuck that’s gonna hurt you mentally retarded since I don’t get paid I won’t go anywhere. Psycho insane retarded. How many more years do you need mentally retarded? when you make your whole family mentally retarded in public
It's a while since I've attended a pre-screening of a movie the last time.
#spiderman
Nice – at Tallinn, Estonia
The main #flatearth grifters out there. All full of shit. – at Maardu, Eesti
#flatearth idiot doesn't understand the atmosphere or how light scattering works. Maybe he should have stayed in school instead of dropping out. – at Maardu, Eesti












