jakeatx/llamAmpere-v0.4-prefill-vs-beellama download history
jakeatx/llamAmpere-v0.4-prefill-vs-beellama is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 95 times (21 in the last 7 days), and 168 times in total. It ranks #156,871 among datasets by monthly downloads.
Prefill speed: llamAmpere v0.4 vs BeeLlama v0.4.7 Prompt processing (prefill) speed of two llama.cpp forks, running the same Qwen3.8-27B GGUF on the same RTX 3090 Ti. We measured prompts from 5,000 to 100,000 tokens in 5,000-token steps, with 3 different prompts at each length. Result
Open jakeatx/llamAmpere-v0.4-prefill-vs-beellama on Hugging Face
Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.