AI Briefing#open-source
The latest AI and developer news about #open-source, with the original source and a short summary.
The latest AI and developer news about #open-source, with the original source and a short summary.
Visual Salamandra, a 7B-parameter multimodal model, has been released with image and video understanding capabilities.
Qwen has released Qwen2.5-VL-32B-Instruct, an open-source multimodal model at 32B scale optimized for performance through reinforcement learning.
Hugging Face submitted a response to the White House's AI Action Plan emphasizing the core role of open source and open science.
NVIDIA has released the Cosmos Transfer model, a large-scale dataset, and Isaac GR00T N1 to advance robotics and autonomous driving.
Mistral released Mistral Small 3.1, an open-source model with multimodal capabilities and a 128k context.
Kanana Flag demonstrated overwhelming performance on Korean benchmarks while reducing training costs by over 50% compared to third-party models.
Qwen has released **QwQ-Max-Preview**, which enhances reasoning and coding performance, and has previewed future app launches and open-sourcing of small models.
It introduces a 3-stage dataset building pipeline and open-source toolkit for training video generation models.
HuggingFace has released an open-source framework to reproduce OpenAI's Deep Research feature.
It summarized the major technical turning points and open-source model achievements in AI image and video generation in 2024.
Mistral AI has released Mistral Small 3, a 24B-parameter model optimized for low-latency performance.
HuggingFace has launched the Open-R1 project to reproduce DeepSeek-R1's training process.
The Diffusers team has summarized the technical status of open video generation models and their future support plans.
Qwen has released the Qwen2.5-1M series, open-source models supporting up to 1 million tokens of context length.
Hugging Face and Entalpic have unveiled LeMaterial, an open source project to accelerate materials research, along with the unified dataset LeMat-Bulk.
The Hugging Face community has released an open-source preference dataset for training text-to-image generation models.
This explains the key provisions of the EU AI Act and what open source developers need to comply with, along with the relevant tools.
HuggingFace has released SmolVLM, a 2B-scale open-source vision language model that maximizes memory efficiency.
The Hugging Face Hub provides various features for data sharing, including large-scale dataset hosting, a data viewer, and third-party library support.
The Qwen2.5-Coder series, featuring GPT-4o-level coding performance, has been released as open source in a variety of model sizes.
Gradio 5 has officially launched with major improvements to performance, design, and real-time streaming capabilities.
SK Planet adopted the open-source Trino as its company-wide data analytics engine, resolving data ingestion delays and building a real-time analytics environment.
Alibaba unveiled the Qwen2.5 model series, trained on 18 trillion tokens, delivering powerful performance.
Alibaba unveiled the Qwen2.5 series, trained on 18 trillion tokens to substantially boost performance.