Process long prompts 3x faster with close to full prefill quality
AI & ML interests
LLM speculative inference, built for specific consumer hardware & heterogeneous computing
Recent Activity
View all activity
models 19
Lucebox/Qwen3.5-9B-GGUF
9B • Updated • 80
Lucebox/DeepSeek-V4.1-Flash-ROCMFP23-GGUF
755B • Updated • 22.3k • 10
Lucebox/DeepSeek-V4.1-Flash-1017-ROCMFP-GGUF
755B • Updated • 28.8k • 10
Lucebox/pflash-scorer-head
2.62M • Updated • 58
Lucebox/pflash-request-probe
265k • Updated • 54
Lucebox/pflash-segment-probe
265k • Updated • 71
Lucebox/Qwen3.8-Flash-Next-GGUF
177B • Updated • 212 • 2
Lucebox/DeepSeek-V4.1-Flash-DSpark-GGUF
14B • Updated • 1.86k • 1
Lucebox/Qwen3.8-27B-IQ4_XS-fast-GGUF
Image-Text-to-Text • 27B • Updated • 551 • 3
Lucebox/DeepSeek-V4-Flash-0731-ROCmFP3
284B • Updated • 2.53k • 27
datasets 0
None public yet