Meta released MobileLLM – 125M, 350M, 600M, 1B model checkpoints! Notes on the release: Depth vs. Width: Contrary to the scaling law (Kaplan et al., 2020), depth is more critical than width for small LLMs, enhancing abstract concept capture and final performance Embedding
@reach_vb
-
Hugging Face Releases AutoTrain Advanced Open Source Code
By
–
And the code lives free for anyone to use too! https://
github.com/huggingface/au
totrain-advanced
… -
The Power and Possibilities of Open Source Software
By
–
The beauty of open source is that you can do all of that and more.
-
Community Praise for Open Source and Scientific Contributions
By
–
Class apart! – I love all that y’all do for open source & science!
-
MiniOmni 2 Development and Production Pipeline Improvements
By
–
Mostly hoping for MiniOmni 2 and likes to get better before putting time into production pipelines for them
-
Choosing TTS Engines for Lightweight GPU Hardware Deployment
By
–
It depends on the hardware you choose to deploy the TTS engine. For lightweight GPUs I’d recommend using MeloTTS
-
AI Voice Synthesis: Reference Speaker Audio Persona Replication
By
–
You just pass a reference speaker audio and it synthesises audio in that persona.
-
Scraping Permissive Data for Better AI Training
By
–
Agreed! Time to scrape permissive and better data
-
Real-time API implementation deployable on L4 under one dollar
By
–
Emphasis on cascaded. We do have real-time API based implementation which you can deploy on a L4 at less than 1$ an hour.