Models & Labs
Hugging Face Migrates Repositories to Xet Storage
Hugging Face Bloghigh confidence
Why it matters
- →Xet storage offers more efficient data handling through content-defined chunking.
- →The migration demonstrates Hugging Face's commitment to improving AI development infrastructure.
- →This shift could significantly enhance collaboration and iteration on large datasets.
Hugging Face has migrated its first model and dataset repositories from LFS to Xet storage, marking a significant advancement in storage efficiency. The new system uses content-defined chunking to deduplicate data at the byte level, allowing for faster and more efficient uploads. This migration has already shifted 6% of the Hub's download traffic to Xet, validating its capability to handle large-scale data transfers. The transition is part of Hugging Face's ongoing efforts to improve collaboration and iteration for AI developers working with massive datasets.
Read originalMore from Hugging Face Blog
© Hugging Face BlogModels & Labsmodels
Hugging Face Launches olmo-eval for LLM Development
Hugging Face has introduced olmo-eval, a new evaluation workbench designed to streamline the iterative process of developing large language models (LLMs). Building on the Open Language Model Evaluation Standard (OLMES), olmo-eval offers enhanced flexibility and modularity, allowing developers to easily configure and run benchmarks across model checkpoints. Unlike traditional evaluation tools, olmo-eval supports agentic and multi-turn evaluations, providing a more nuanced analysis of model improvements. This tool is particularly useful for developers who need to quickly assess the impact of changes in data, architecture, or hyperparameters during the model development cycle.
Hugging Face Blog
© Hugging Face BlogResearchresearch
Profiling PyTorch: From nn.Linear to Fused MLP
Hugging Face's blog post dives into the profiling of PyTorch operations, focusing on the shift from basic matrix operations to using nn.Linear and constructing a Multilayer Perceptron (MLP). The article reveals how nn.Linear manages operations by integrating bias addition into the matrix multiplication kernel, effectively reducing overhead. It also examines the limited impact of torch.compile on single operations, pointing out its potential in more complex scenarios. These insights are crucial for developers aiming to optimize deep learning models on GPUs, as they provide a deeper understanding of how to maximize performance and efficiency.
Hugging Face Blog
More in Models & Labs
Models & Labsmodels
vLLM v0.23.0 Release Enhances Model Support
The vLLM v0.23.0 release marks a significant step forward with enhancements across various components. DeepSeek-V4 has been optimized further, decoupling its metadata from previous versions and adding new attention kernels. Model Runner V2 now supports more dense models by default, improving performance for Llama and Mistral. The Rust frontend has matured with new endpoints and tool parsers, while compatibility with Transformers v5 ensures broader model support. These updates collectively enhance the robustness and versatility of vLLM, making it a more powerful tool for developers working with large language models.
vLLM Releases
Models & Labsmodels
Llama.cpp b9626 Release Adds Cohere2-MoE Support
The latest b9626 release of llama.cpp introduces architectural support for the cohere2-MoE model, marking a significant update for developers working with this model. This release also includes various technical improvements such as the removal of redundant checks and enhancements in tensor handling, which streamline the model's performance. By adding cohere2moe to the Llama Model Saver supported list, the update broadens the toolkit available for AI practitioners. While these changes may seem incremental, they collectively enhance the robustness and flexibility of llama.cpp, making it a more versatile tool for AI development.
llama.cpp Releases
Models & Labsmodels
llama.cpp b9627 Release Expands Platform Support
The b9627 release of llama.cpp continues to enhance its platform reach, though it doesn't introduce any groundbreaking features. This update includes support for a wide array of systems, from macOS and iOS to various Linux distributions and Windows configurations, including CUDA and Vulkan support. Notably, the release maintains its focus on making llama.cpp a versatile tool across different hardware setups, but it doesn't introduce new model architectures or quantization methods. This iteration is more about solidifying its presence across multiple operating systems rather than introducing novel capabilities.
llama.cpp Releases