AI Briefing#huggingface
The latest AI and developer news about #huggingface, with the original source and a short summary.
The latest AI and developer news about #huggingface, with the original source and a short summary.
HuggingFace has released SmolVLM, a 2B-scale open-source vision language model that maximizes memory efficiency.
Hugging Face is introducing a new CAS-based upload/download architecture for efficient transfer of large models and datasets.
Hugging Face and LLM-jp collaborated to release an open leaderboard evaluating the performance of Japanese LLMs.
Hugging Face is introducing Content-Defined Chunking (CDC) technology to overcome the limitations of Git LFS, improving storage and transfer efficiency.
Judge Arena, a benchmarking platform that compares and ranks LLMs' evaluation capabilities, has launched.
The Hugging Face Hub provides various features for data sharing, including large-scale dataset hosting, a data viewer, and third-party library support.
An integration feature has been released that allows users to search for Hugging Face models and insert code directly within the PyCharm IDE.
Argilla 2.4 has been released, enabling data labeling and feedback collection by importing datasets from the Hugging Face Hub without any code.
A technology has been developed that can speed up LLM inference using any small model, even if the tokenizer is different.
CinePile 2.0, a video QA dataset with improved performance through adversarial refinement techniques, has been released.
Hugging Face has launched HUGS, an inference service that makes it easy to deploy open models on your own infrastructure.
Google DeepMind and Hugging Face have released SynthID Text, which embeds and detects watermarks in AI-generated text.
Hugging Face has released a guide on how to deploy Speech-to-Speech (S2S), a complex pipeline, to Inference Endpoints.
Outlines-core 0.1.0, a Rust port of Outlines' core algorithm for supporting structured output in LLMs, has been released.
Hugging Face's Diffusers library officially supports the Stable Diffusion 3.5 Large model.
Transformers.js v3 has been released, maximizing browser-based AI inference performance through WebGPU support.
Hugging Face is enhancing its model security scanning capabilities by integrating Protect AI's Guardian.
Through Keras Hub, Llama 3.2 models can be easily used across JAX, PyTorch, and TensorFlow backends.
HuggingFace fixed an issue in the Transformers library where loss function calculation during Gradient Accumulation was mathematically inaccurate.
Gradio has released Gradio 5, fixing all vulnerabilities discovered through a security audit by Trail of Bits.
It explains how to use Dask to scale parallel model inference and data preprocessing on Hugging Face's large-scale datasets.
Gradio 5 has officially launched with major improvements to performance, design, and real-time streaming capabilities.
Dynamic speculative decoding, developed by Intel Labs and Hugging Face, has been applied to the Transformers library, boosting generation speed by up to 2.7x.
Hugging Face analyzed deduplication issues in Parquet files for large-scale dataset efficiency and proposed improvements.