OPEN SOURCE - AI Dynamics

Meta AI Scales Speech Technology to Over 1000 Languages

By

–

21 June 2023 16h19

Meta AI released Scaling Speech Technology to 1,000+ Languages

demo: https://t.co/1yMek3Uzbl

Expanding the language coverage of speech technology has the potential to improve access to information for many more people. However, current speech technology is restricted to about… pic.twitter.com/BHmA7J5YnW
— AK (@_akhaliq) 21 juin 2023

Meta AI released Scaling Speech Technology to 1,000+ Languages demo: https://
huggingface.co/spaces/mms-met
a/MMS
… Expanding the language coverage of speech technology has the potential to improve access to information for many more people. However, current speech technology is restricted to about

→ View original post on X — @_akhaliq

21 June 2023

MMS Integration in Transformers: Easy Inference and Fine-Tuning

By

@reach_vb

–

21 June 2023 16h01

5 / 5 We've integrated MMS into Transformers making the models very easy to use for both inference and fine-tuning. Demo https://
huggingface.co/spaces/faceboo
k/MMS
…
Docs https://
huggingface.co/docs/transform
ers/main/en/model_doc/mms
…
Fine-Tuning

→ View original post on X — @reach_vb

21 June 2023

Meta’s Multilingual Speech Model Democratizes AI Access Globally

By

@reach_vb

–

21 June 2023 16h00

Meta AI's recently released "Massively Multilingual Speech" (MMS) model is a huge step forward towards democratizing
Speech to every corner of the globe. In addition, it might also play a significant role in preserving global linguistic diversity. https://
huggingface.co/papers/2305.13
516
…

→ View original post on X — @reach_vb

21 June 2023

Free Midjourney Alternative: AlphaCTR AI Model Launch

By

@waitin4agi_

–

21 June 2023 9h09

For those keen on a free MJ alternative, check out our model: http://
AlphaCTR.com

→ View original post on X — @waitin4agi_

21 June 2023

Phi-1: Efficient 1.3B Parameter Language Model for Code

By

@_akhaliq

–

21 June 2023 5h33

Textbooks Are All You Need paper page: https://
huggingface.co/papers/2306.11
644
… introduce phi-1, a new large language model for code, with significantly smaller size than competing models: phi-1 is a Transformer-based model with 1.3B parameters, trained for 4 days on 8 A100s, using a selection

→ View original post on X — @_akhaliq

21 June 2023

Open Source Language Models Race Gains Expert Attention

By

@mattlynley

–

21 June 2023 5h26

Well there we go. I’d heard it was at least 4 experts but more mechanics on how it works out here. (Prev reported on it last week: https://
supervised.news/p/the-race-for
-open-source-language
…)

→ View original post on X — @mattlynley

21 June 2023

RepoFusion: Training Code Models to Understand Repository Context

By

@_akhaliq

–

21 June 2023 5h05

RepoFusion: Training Code Models to Understand Your Repository paper page: https://
huggingface.co/papers/2306.10
998
… Despite the huge success of Large Language Models (LLMs) in coding assistants like GitHub Copilot, these models struggle to understand the context present in the repository (e.g.,

→ View original post on X — @_akhaliq

21 June 2023

Daft: A Python DataFrame for Complex Data

By

@miketamir

–

21 June 2023 3h01

GitHub – Eventual-Inc/Daft: The Python DataFrame for Complex Data
https://bit.ly/446lC24 #AI #MachineLearning #DeepLearning #LLMs #DataScience

→ View original post on X — @miketamir

21 June 2023

Tinygrad and Open Source ML Frameworks Competition Excitement

By

@soumithchintala

–

20 June 2023 23h44

I really enjoyed this podcast from @realGeorgeHotz (the transcripts make it much faster to grok). I'm genuinely excited about Tinygrad (and ggml and JAX). When there's competition that is exploring a new direction or philosophy, its exciting and fun and there's nothing more

→ View original post on X — @soumithchintala

20 June 2023