The v0.18.2rc0 release includes a fix for handling the max_pixels parameter in the PaddleOCR-VL image processor across transformations.
Read originalThe v0.19.0rc0 release introduces a feature for CPU key-value cache offloading, enhancing performance. This update was signed off by Yifan Qiao.
The latest release of Llama.cpp introduces new Vulkan functions for tensor manipulation and updates across multiple platforms.
The latest release of llama.cpp includes support for various operating systems and architectures, including macOS, Linux, Android, and Windows. This update enhances compatibility for developers working across different environments.
The v0.19.0rc1 release includes a bug fix that restricts TRTLLM attention to SM100, addressing issues with GB300 (SM103).