Gabriel F. Rosalino

@maximelabonne @liquidai @iclr_conf Hey, big move! Best of luck for you guys. If you intend to hire SWEs or MLEs, let me know

🎉 Big news! I've joined @liquidai, an MIT spin-off, where I'm leading the efforts to fine-tune LLMs. They've got big plans and serious compute power, so I'm excited to see what we can accomplish :) If you want to meet in person, I'll be at our social event at @iclr_conf in Vienna next week: docs.google.com/forms/d/e/1FAI…

@omarsar0 Do you think it’s possible to achieve multi modality through models merging? Do you know any known methods for fine tuning multi modal open source algorithms?

RAG vs. Fine-Tuning Cool report discussing the tradeoff between RAG and fine-tuning when using LLMs like Llama 2 and GPT-4. It performs a detailed analysis and highlights insights when applying the pipelines on an agricultural dataset. Here is a figure showing the pipeline used in this study: Here is a summary of the comparison between RAG and fine-tuning results: Findings: The authors observe that there is an "accuracy increase of over 6 p.p. when fine-tuning the model and this is cumulative with RAG, which increases accuracy by 5 p.p. further." They also "demonstrate that the fine-tuned model leverages information from across geographies to answer specific questions, increasing answer similarity from 47% to 72%." RAG is effective where data is contextually relevant such as interpretation of farm data. However, it might significantly increase the prompt size and become harder to steer. Fine-tuning, on the other hand, could be tuned for brevity and can incur less cost (i.e., necessitates minimal input token size) when dealing with large datasets. The challenge is the initial cost and effort required to fine-tune models on new data. Overall, the suitability of each approach depends on the specific application, the nature and size of the data, and available resources for model development. As suggested by many other reports, there is also the possibility of combining the two approaches. I also agree with the authors that it would be interesting to combine structured information from PDFs with images and captions to enable multi-modal fine-tuning opportunities.

@gympasshelp I'm locked out of my account. There's no answer to my emails, and I can't login to talk with support in the chat. Help!

In 2017 @sama wrote a blog post about what makes the ideal cofounder. It was titled “Greg” @gdb Very cool. blog.samaltman.com/greg

If you're having problems with your Continuous Deployment or Continuous Integration pipelines, and you use PHP stack, this might be the problem (SSL wildcard certificate expired for pecl.php.net). Quick dirty fix: download the deps tarball packages and use the files!

@levelsio Largest country of people pleasers

Why are Brazilians so nice?

Stop saying: AI will replace humans. Start saying: humans who know how to use AI at work will replace those who don’t.

Abstraction is the best & worst part of programming. Amazing abstractions give the power to easily build something great; bad abstractions mean not getting anything done. Worth waiting much later to get started, and then much more effort iterating on them, than most people think.

Thanks Parcels for spending 5,522 minutes with me in 2022. I couldn’t stop listening to Bemyself - from Hansa Studios, Berlin. #SpotifyWrapped open.spotify.com/artist/3oKRxps…

🐶 oi humanos, meu nome é Bianca, mas pode me chamar de Bia

Mais comum do que gostaríamos AHAHAHA

Agradeço a você, Parcels, por passar 1,484 minutos comigo no @Spotify esse ano. Pra mim, não teve artista como você em 2020. #RetrospectivaSpotify open.spotify.com/artist/3oKRxps…

The most powerful paradoxes of life:

Ou a organização inteira adota a gestão ágil como modus operandi, ou a alinhamento/comunicação entre TI e os outros departamentos vai ser extremamente inconsistente. Pra não falar no estresse que vai gerar...

@o_gabsferreira Obrigado por compartilhar!!

Entender padrões de arquitetura de software quando você é iniciante tá longe de ser fácil. Esse artigo tenta resumir os principais padrões existentes em 5 minutos de leitura: orkhanscience.medium.com/software-archi…

@rslulanik Tenho ouvido muitos amigos devs dizendo que é possível viver bem com 8gb de ram no M1. Minha dúvida é: será que guenta IDE da JetBrains + Docker + VPN? Vc usa Docker?

@rslulanik Tá usando M1 com 8gb de ram ou 16gb?

Utilidade pública.

"First, solve the problem. Then, write the code." — John Johnson

Hey, @graneromudanca: Dá uma olhada nessa reclamação contra Granero Transportes no @ReclameAQUI reclameaqui.com.br/granero-transp…