From the source

Other stories

AI lab announcements, news reports and podcasts.

Feed Chatter About Privacy Connect RSS

© 2026 forck.live

From the source

Other stories

Vision Language Model Alignment in TRL ⚡️ — forck

Copy▾✕ Close copy menu

Hugging Face· 7 Aug

Vision Language Model Alignment in TRL ⚡️

Developer tool

Image published with the post by Hugging Face — Image from huggingface.co

Hugging Face announces new alignment methods for Vision Language Models in TRL: Mixed Preference Optimization (MPO), Group Relative Policy Optimization (GRPO), Group Sequence Policy Optimization (GSPO), plus extensions for RLOO and Online DPO, and native SFT support for VLMs.

From the source
Here’s what’s new in TRL: Mixed Preference Optimization (MPO) Group Relative Policy Optimization (GRPO) Group Sequence Policy Optimization (GSPO) (a variant of GRPO)
huggingface.co

Other stories

Hugging Face· 15 OctModel update

Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint

Granite Switch 4.1 models combine chat, decisions, and speech-to-text from one endpoint.

Hugging Face· 15 OctModel update

Apertus 1.5 Release Notes

Apertus 1.5 adds multimodal understanding and thinking mode

Announced: 07 Aug 2025
Detected: 03 Aug 2026 21:12 UTC (+360d after)

ModelsIDEFICS2

Visit source ↗huggingface.co

Hugging Face· 15 OctOpen-weight release

Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta

Meta releases open-weight multimodal MoE model with 816B parameters

Hugging Face· 15 OctOpen-weight release

Welcome FLUX.3 Image Klein in Diffusers

Open image model adds layout control via bounding boxes

Hugging Face· 8 Oct

The model that didn't exist, so you made it yourself

A user creates six custom models via ML Intern in days

Hugging Face· 2 OctResearch paper

AutoSynthData: Generating Training Data for Enterprise Agents

Turning model failures into training data for enterprise agents

Most covered

Meta Muse Wonder: A Frontier Model For Personal & Professional AI Agents
SiliconANGLE: “Meta releases Muse Wonder AI models built to run personal and professional agents”
Reporting · 29 outlets
Meta· 12 Oct
Muse Wonder: A Frontier Model For Personal & Professional AI Agents
Meta Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date
SiliconANGLE: “Meta releases 816B-parameter Muse Spark Open for coding and agent work”
Reporting · 27 outlets
Image from research.meta.ai
Meta· 15 Oct
Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date
Anthropic An update on independent evaluators
CNBC: “Anthropic taps nonprofits METR and SecureBio to help assess risks and safety practices”
Reporting · 9 outlets
Image from anthropic.com
Anthropic· 15 Oct
An update on independent evaluators
NVIDIA Be There for the Beginning: Celebrate the Launch of NVIDIA RTX Spark
The Verge: “The best reason to care about Nvidia’s RTX Spark: ‘everything’ will run on Arm”
Reporting · 8 outlets
Image from blogs.nvidia.com
NVIDIA· 15 Oct
Be There for the Beginning: Celebrate the Launch of NVIDIA RTX Spark
xAI Introducing Grok Voice 3.0
SiliconANGLE: “Grok Voice 3.0 from SpaceXSI outscores OpenAI and Google on speech-to-speech index”
Reporting · 8 outlets
Image from x.ai
xAI· 14 Oct
Introducing Grok Voice 3.0