The 🇨🇿 BenCzechMark is the first and most comprehensive evaluation suite for assessing the abilities of Large Language Models (LLMs) in the Czech language. It aims to test how well LLMs can: Reason and perform complex tasks in Czech. Generate and verify grammatically and semantically correct Czech. Extract information and store knowledge by answering questions about Czech culture and Czech-related facts. Do what language models were originally trained for—estimate the probability of Czech texts. To achieve this, we’ve sourced 50 tasks […]

In the early 15th century, Zheng He (also known as Chong Ho), a Chinese mariner and explorer during the early Ming Dynasty, led seven major naval expeditions, known as the “Voyages to the Western Oceans”. His journey traced a path that went through Southeast Asia, the Middle East and then reached out to Africa. It was a bold move by China to    

Finding the best LLM models for finance use cases The growing complexity of financial language models (LLMs) necessitates evaluations that go beyond general NLP benchmarks. While traditional leaderboards focus on broader NLP tasks like translation or summarization, they often fall short in addressing the specific needs of the finance industry. Financial tasks, such as predicting stock movements, assessing credit risks, and extracting information from financial reports, present unique challenges that require models with specialized skills. This is why we decided […]

The Xet team at Hugging Face is working on improving the efficiency of the Hub’s storage architecture to make it easier and quicker for users to store and update data and models. As Hugging Face hosts nearly 11PB of    

The Hugging Face platform has many datasets and pre-trained models that make using and training state-of-the-art machine learning models increasingly accessible. However, it can be hard to scale AI tasks because AI datasets are often large (100s GBs to TBs) and using Hugging Face transformers for model inference can sometimes be computationally expensive. Dask, a Python library for distributed computing, can handle out-of-core computing (processing data that doesn’t fit in memory) by breaking datasets into manageable chunks. This makes it […]

We’ve been hard at work over the past few months, and we are excited to now announce the stable release of Gradio 5.  With Gradio 5, developers can build production-ready machine learning web applications that are performant, scalable, beautifully designed, accessible, and follow best web security practices, all in a few lines of Python. To give Gradio 5 a spin, simply type    

We audited Gradio 5 so that your machine learning apps are safe! In the past few years, Gradio (>6 million monthly Pypi installs) has become the default way to build machine learning web applications in Python. In just a    

AMD has just unveiled its 5th generation of server-grade EPYC CPU based on Zen5 architecture – also known as Turin. It provides a significant boost in performance, especially with a higher number of core count reaching up to 192    

Our friends at Unsloth shared an issue regarding gradient accumulation yesterday that is affecting the transformers Trainer. The initial report comes from @bnjmn_marie (kudos to him!). Gradient accumulation is supposed to be mathematically equivalent to full batch training; however, losses did not match between training runs where the setting was toggled on and off. Where does it stem from? Inside the modeling code of each model, transformers offers a “default” loss function that’s the most typically    

This is going to be the shortest blog post ever. Question: Llama 3.2 landed two weeks ago on Hugging Face / Transformers. When will it be available in Keras? Answer: It has been working from day 1 😀. There is nothing to wait for. Yes, Keras Llama3 can be loaded from any standard (i.e. safetensors) Hugging