AI BriefingHuggingFace Blog news
The latest AI and developer news from HuggingFace Blog, with short summaries.
Visit website →TRL v1.14 Adds LoRA Async GRPO Support Without NCCL
TRL v1.14 introduces Async GRPO functionality that synchronizes LoRA adapters using storage buckets and proxies instead of NCCL.
Gradio Releases Workflow1111, Replicating Key A1111 Features
Gradio has released 'Workflow1111', which replicates the key features of AUTOMATIC1111 in a single workflow canvas.
IBM Releases Open-Source Time Series Model PatchTST-FM-r2
IBM has released Granite PatchTST-FM-r2, a time series foundation model licensed under Apache 2.0.
Multiverse Computing Reveals Safety Boundary Control Technique to Reduce LLM Over-Refusal
Multiverse Computing reduced over-refusal from 74% to 4% using a 'Narrow-boundary safety' technique that refuses only specific subsets rather than entire topics.
NeoMME Releases Single Encoder
NeoMME, a multilingual multimodal encoder that processes text and images with a single Transformer without a vision tower, has been released.
Hugging Face Releases 'funes', a Local Memory Layer for Coding Agents
Hugging Face has released 'funes', a local open-source memory tool that preserves cross-session reasoning for coding agents like Claude Code.
Training a Coding Model to Paint Watercolors Using TRL and OpenEnv
An open-source reproduction project has been released for a watercolor-generating coding model trained with human aesthetic preferences as the reward function, utilizing TRL and OpenEnv.
IBM and Confluent Release Real-Time Time-Series AI Model
IBM and Confluent have released the IBM Granite time-series foundation model in Early Access, operating within real-time streaming environments based on Apache Flink.
AI2 Releases Benchmark Analysis Method
AI2 has released BenchMIRT, a multidimensional IRT-based tool that analyzes which specific capabilities individual questions in LLM benchmarks actually measure.
Hugging Face Releases WebGPU Kernels
Hugging Face has released 207 optimized WebGPU kernels and a loader library to enhance in-browser AI inference performance.
GLM-5.3 Open Weights Released
Z.ai released the GLM-5.3 open-weight model, featuring a 50% improvement in coding performance over GLM-5.2 and enhanced cybersecurity capabilities.
HF Introduces Hindi ASR Evaluation
Hugging Face has released Monsoon, a speech recognition evaluation dataset for Hindi and Indian English, to support Global South languages.
Sentence Transformers v6.0 Adds Support for Multi-Vector Training
Sentence Transformers v6.0 introduces multi-vector model training capabilities, significantly improving domain-specific search performance.
IBM Releases Granite 4.2 Reasoning LLMs
IBM has released three Granite 4.2 reasoning LLMs featuring 512K context and agentic RL.
4-bit LLMs Surpass Original Model Performance
The QAH technique is released, recovering performance higher than the original model after structural compression and 4-bit quantization.
Gradio Releases gr.Workflow Feature
Gradio has launched the gr.Workflow feature, which supports building visual node-based AI pipelines and deploying APIs.
Building Hybrid Search on HF Infrastructure for Papers with Code
Revealed the architecture and design decisions for building Papers with Code's hybrid search system using Hugging Face infrastructure.
ASR Models Memorize Benchmark Answers
A study quantifies the phenomenon of ASR models inflating scores by learning benchmark patterns rather than actual speech.
LFM2.5 Inference Accelerated 3.2x with DSpark
Applying DSpark speculative decoding to LFM2.5 models improves GPU and on-device inference speed by up to 3.2x.
Liquid AI Releases LFM2.5 Q4_0
Liquid AI has released the LFM2.5 Q4_0 models, which minimize quality degradation through Quantization-Aware Distillation (QAD).