ggml-org/llama.cpp
2 storiesEvery ggml-org/llama.cpp story collected by The AI Daily, 2 so far, newest first, refreshed hourly.
Related topics
A PR to ggml-org/llama.cpp adds the Maple 20B-A1B ternary MoE architecture by AlexGabbia, runnable on CPU and visible as deepgrove/maple-preview on Hugging Face. The change extends open-source options for users with low VRAM.
A ggml-org/llama.cpp PR by pwilkin adds Flash Attention tuning for gfx1201 (RDNA4) and some RDNA3 GPUs (e.g., RX 9060 XT, 8060S), with detailed benchmarks showing improved large-context performance.
That is everything
The A.I. Daily
llama.cpp adds Maple 20B-A1B ternary MoE (CPU)