Tagged "fp8-quantization"
3 articles tagged fp8-quantization, 13 February 2026 to 11 November 2026. Newest first.
-
llama.cpp Adds FP8 Quantisation Support for CUDA
llama.cpp releases major update adding FP8 (8-bit floating point) quantisation support with CUDA optimization, reducing model memory footprint while maintaining inference quality.
-
llama.cpp v0.9.0 Adds OCP FP8 Quantization and Flash Attention Improvements
llama.cpp releases v0.9.0 introducing OCP FP8 E4M3 quantization support and a new flash_attn_ext_rows operation for efficient multi-sequence inference. The release also adds support for four new model architectures and audio generation capabilities.
-
Ring-1T-2.5 Released with SOTA Deep Thinking Performance
inclusionAI releases Ring-1T-2.5 in FP8 format, claiming state-of-the-art performance on deep thinking tasks with optimized quantization for local deployment.