AI BriefingHuggingFace Blog news
The latest AI and developer news from HuggingFace Blog, with short summaries.
Visit website →Analyzing the Cause of MMLU Benchmark Score Discrepancies
This analyzes the technical causes behind the differences in MMLU benchmark scores between the Open LLM Leaderboard and model papers.
Hugging Face Supports Panel Templates
With Panel templates on Hugging Face Spaces, you can easily build and deploy data dashboards and apps.
Hugging Face Submits Comments on US NTIA's AI Accountability Policy
Hugging Face submitted a response to the US NTIA's request for comments on AI accountability, emphasizing transparency and multidisciplinary engagement.
Maximizing Low-Resource Language ASR Performance with MMS Adapters
This introduces a method for efficiently boosting ASR performance for low-resource languages by fine-tuning Meta's MMS model using an adapter approach.
Time Series Forecasting: Transformers Overwhelm Linear Models
The transformer-based model Autoformer has proven superior to the simple linear model DLinear in time series forecasting performance.
HuggingFace Announces New Content Policy
HuggingFace has announced a new content policy for the responsible development of AI models and datasets.
Hugging Face Spaces Now Supports Livebook Deployment
Elixir-based Livebook notebooks can now be deployed as apps on Hugging Face Spaces.
Core ML-based Stable Diffusion Acceleration
Through Apple's Core ML optimizations and 6-bit palettization, Stable Diffusion can run faster and more efficiently on Apple devices.
Hugging Face-AMD Partner to Accelerate AI Models
Hugging Face and AMD have formed a partnership to optimize transformer model performance on AMD CPUs and GPUs.
Hugging Face Expands LLM Evaluation System
Hugging Face has expanded the evaluation methodology of the Open LLM Leaderboard to verify the reliability of LLMs for data labeling.
Hugging Face Adds DuckDB Support for Dataset Analysis
A new feature has been added that allows datasets on Hugging Face Hub to be analyzed directly with SQL via DuckDB.
Hugging Face Begins Supporting fastText Models
Meta AI's fastText models and word vectors for 157 languages have been integrated into the Hugging Face Hub.
TII Releases Open-Source LLM 'Falcon'
The high-performance open-source LLM Falcon series developed by TII has been released into the Hugging Face ecosystem.
BERTopic Adds Hugging Face Hub Integration Support
An integration feature has been added that allows BERTopic models to be saved to and loaded directly from the Hugging Face Hub.
LLM Inference Container for SageMaker Released
Hugging Face has launched a TGI-based LLM inference container dedicated to Amazon SageMaker.
5.1x Faster Stable Diffusion on Intel CPUs
This presents an optimization method using NNCF and Optimum that boosts Stable Diffusion inference speed on Intel CPUs by 5.1x.
Instant Deployment of Hugging Face Models on Azure
Hugging Face and Microsoft have partnered to integrate the Hugging Face model catalog into Azure Machine Learning.
4-bit Quantization and QLoRA Released
Thanks to 4-bit quantization and QLoRA technology, large language models can now be efficiently fine-tuned even with limited memory.
Stable Diffusion Performs Image Editing Instructions
This introduces an instruction-tuning methodology that enables Stable Diffusion to understand and carry out users' image editing instructions.
Safetensors Security Verification and Transition to Default Format
The Safetensors library is establishing itself as the default format for model storage after proving its safety through an external security audit.