Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
projectlosangeles 
posted an update 13 days ago
GGUFGuy

But can i run it on colab t4?

·
projectlosangeles

@GGUFGuy

Not on T4 since it seem to require at least 22GB VRAM. But you can run it on other Google Colab GPUs like A100.

Alternatively, you can duplicate official space and run it here on HF hardware with enough VRAM or even ZeroGPU.

Hope this helps :)
projectlosangeles

🤗Here is the select, best-of-the-best version of my HeartMuLa output samples album to save ya all time and to make listening experience better!!! 🤗

❤️https://soundcloud.com/aleksandr-sigalov-61/sets/heartmula-select❤️

Overall, my experience with HeartMuLa was very positive and fun! It certainly beats DiffRhythm and crappy ACE Step in both song controllability and especially sound fidelity/quality.

The model works very fast (1.5 min per 4 min song with torch compile) and uses only 22GB VRAM.

Requirements, setup and use is also very simple and nice.

Most sincerely,

Alex
asigalov61

🤗Thank ya to all who liked and otherwise appreciated my post!!! 🤗

Sincerely,

Alex
John6666
•
edited 7 days ago

oh, link is broken due to 🤗 mark..
https://huggingface.co/HeartMuLa/HeartMuLa-oss-3B-happy-new-year

Edit:
done.

·
projectlosangeles

Thank you. Fixed :)
In this post