ggml/llama.cpp releases: v0.4.0
Sentiment: neutral
TL;DR
Version 0.4.0 of llama.cpp was released, adding support for Qwen3.8-Flash-Next and Nemotron-3-Puzzle models, along with several new features like on-demand tensor reading and video input options, making it more versatile for AI language tasks. This update is significant as it enhances the model's capabilities and flexibility, catering to a broader range of applications in natural language processing.
Detailed Summary
The ggml/llama.cpp project released version 0.4.0, which introduces support for Qwen3.8-Flash-Next and Nemotron-3-Puzzle models, along with enhancements like on-demand tensor reading and video input options. This update also includes improvements to the ggml library, specifically major sparse flash attention and RDMA work. The broader impact includes expanded model compatibility and enhanced functionality for developers working with large language models.
Key Points
- • Adds initial Qwen3.8-Flash-Next and Nemotron-3-Puzzle support
- • Includes on-demand tensor reading feature
- • Introduces per-slot server context limits