4 26 101

João Palmeiro

joaompalmeiro

AI & ML interests

None yet

Recent Activity

reacted to Akhil-Theerthala's post with 👍 about 17 hours ago

Is it better to show a model too many images once (Diversity), or extract as much information from a small set of images? I have always wanted to do an ablation study on this and recently I got the chance to do exactly that. Why? In applied domains like robotics, manufacturing, or banking, we rarely have the luxury of internet-scale diverse image datasets. We are often "Data Poor" in terms of diversity but "Data Rich" in depth. The takeaway? Density is efficient for facts but dangerous for reasoning (logical collapse) if you don't have larger scale data. More details: https://huggingface.co/blog/Akhil-Theerthala/diversity-density-for-vision-language-models

upvoted a collection about 17 hours ago

Density-Vs-Diversity-Blogpost

upvoted an article 1 day ago

Diversity Vs Density: A data strategy comparison for fine-tuning VLMs

View all activity

Organizations

upvoted a collection about 17 hours ago

Density-Vs-Diversity-Blogpost

Collection

The collection contains the artefacts used to do the analysis for the blogpost: Diversity Vs Density: A strategy comparison for fine-tuning VLMs • 7 items • Updated 1 day ago • 2

upvoted an article 1 day ago

Article

Diversity Vs Density: A data strategy comparison for fine-tuning VLMs

2 days ago

•

upvoted a collection about 1 month ago

Ministral 3

Collection

A collection of edge models, with Base, Instruct and Reasoning variants, in 3 different sizes: 3B, 8B and 14B. All with vision capabilities. • 9 items • Updated Dec 2, 2025 • 138

upvoted a collection 3 months ago

Qwen3-VL

Collection

37 items • Updated 8 days ago • 559

upvoted an article 3 months ago

Article

When Does Reasoning Matter? Unpacking the Contribution of Reasoning to LLM Performance

Sep 30, 2025

•

upvoted a paper 3 months ago

Benchmark It Yourself (BIY): Preparing a Dataset and Benchmarking AI Models for Scatterplot-Related Tasks

Paper • 2510.06071 • Published Oct 7, 2025 • 1

upvoted an article 4 months ago

Article

Jupyter Agents: training LLMs to reason with notebooks

Sep 10, 2025

•

upvoted an article 5 months ago

Article

Introducing Command A Vision: Multimodal AI built for Business

Jul 31, 2025

•

upvoted 2 articles 6 months ago

Article

Reachy Mini - The Open-Source Robot for Today's and Tomorrow's AI Builders

Jul 9, 2025

•

755

Article

SmolLM3: smol, multilingual, long-context reasoner

Jul 8, 2025

•

744

upvoted a paper 6 months ago

Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens

Paper • 2506.17218 • Published Jun 20, 2025 • 29

upvoted 2 papers 7 months ago

VisionZip: Longer is Better but Not Necessary in Vision Language Models

Paper • 2412.04467 • Published Dec 5, 2024 • 117

SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics

Paper • 2506.01844 • Published Jun 2, 2025 • 147

upvoted an article 7 months ago

Article

Context Is Gold to Find the Gold Passage: Evaluating and Training Contextual Document Embeddings

Jun 2, 2025

•

upvoted a paper 7 months ago

Why Vision Language Models Struggle with Visual Arithmetic? Towards Enhanced Chart and Geometry Understanding

Paper • 2502.11492 • Published Feb 17, 2025 • 2

upvoted an article 8 months ago

Article

Common AI Model Formats

Feb 27, 2025

•

upvoted a paper 8 months ago

ChartMuseum: Testing Visual Reasoning Capabilities of Large Vision-Language Models

Paper • 2505.13444 • Published May 19, 2025 • 17

upvoted 2 articles 8 months ago

Article

Fine-tune Llama 3.1 Ultra-Efficiently with Unsloth

Jul 29, 2024

•

365

Article

Preference Optimization for Vision Language Models

Jul 10, 2024

•

upvoted a paper 8 months ago

Marigold: Affordable Adaptation of Diffusion-Based Image Generators for Image Analysis

Paper • 2505.09358 • Published May 14, 2025 • 26

João Palmeiro

AI & ML interests

Recent Activity

Organizations

joaompalmeiro's activity

Diversity Vs Density: A data strategy comparison for fine-tuning VLMs

When Does Reasoning Matter? Unpacking the Contribution of Reasoning to LLM Performance

Jupyter Agents: training LLMs to reason with notebooks

Introducing Command A Vision: Multimodal AI built for Business

Reachy Mini - The Open-Source Robot for Today's and Tomorrow's AI Builders

SmolLM3: smol, multilingual, long-context reasoner

*Context Is Gold to Find the Gold Passage*: Evaluating and Training Contextual Document Embeddings

Common AI Model Formats

Fine-tune Llama 3.1 Ultra-Efficiently with Unsloth

Preference Optimization for Vision Language Models

Context Is Gold to Find the Gold Passage: Evaluating and Training Contextual Document Embeddings