On Common Sense Reasoning, Closed-book Question Answering, and Reading Comprehension, LLaMA-65B outperforms Chinchilla 70B and PaLM 540B on almost all benchmarks.
4/n
OPEN SOURCE
-
LLaMA-65B Outperforms Chinchilla and PaLM on Reasoning Benchmarks
By
–
-
Open Dataset Approach: LLaMA’s Reproducible Alternative to Chinchilla and GPT-3
By
–
Unlike Chinchilla, PaLM, or GPT-3, we only use datasets publicly available, making our work compatible with open-sourcing and reproducible, while most existing models rely on data which is either not publicly available or undocumented.
2/n -
Meta Releases LLaMA Foundation Models Open Source
By
–
Today we release LLaMA, 4 foundation models ranging from 7B to 65B parameters.
LLaMA-13B outperforms OPT and GPT-3 175B on most benchmarks. LLaMA-65B is competitive with Chinchilla 70B and PaLM 540B.
The weights for all models are open and available at https://
research.facebook.com/publications/l
lama-open-and-efficient-foundation-language-models/
…
1/n -
AI Transparency and Open Source: Addressing Knowledge Gaps
By
–
Don’t want to be a gate-keeper or anything but it’s wild how many people talk about AI and machine learning that they barely understand right now! We need more transparency & openness from builders (research papers/oss) for it not to get out of hands and stay closer to the truth!
-
Running Automatic1111 Locally for AI Image Generation
By
–
Es Automatic1111 en local. Pero debería de funcionar en cualquier versión de Google Colab de Automatic.
-
ControlNet Models Fully Compatible with DreamBooth Trained Models
By
–
Por si teníais dudas, AFIRMATIVOS. Los modelos de ControlNet SÍ son compatibles con los modelos entrenados con DreamBooth
-
Meta AI Newsletter: Open Source Work, Datasets, Community
By
–
A new issue of the Meta AI newsletter is heading to your inbox soon! This edition includes a look at some of our recent open source work, datasets and opportunities to get connected with the AI community! Subscribe
-
Zero-shot Text Classification Models and APIs for NLP Tasks
By
–
Take a look at Hugging Face zero-shot text classification page, many models there: https://
huggingface.co/tasks/zero-sho
t-classification
… You can also try APIs(Open AI, Cohere, etc…). Easier option maybe. -
Hugging Face presence indicator for Twitter AI community
By
–
If you don't have HF in your twitter summary, it's your cue that you haven't tweeted enough about HF https://
huggingface.co/spaces/Tuana/s
hould-i-follow#lighttheme
… -
BLIP-2 Open Source Multimodal Model Capabilities Demonstrated
By
–
Este y más ejemplos de este potente modelo open source, BLIP-2, lo encontraréis en mi último vídeo. https://
youtu.be/VZQJuY-cL8w