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 17
Lucebox/pflash-scorer-head
2.62M • Updated • 9
Lucebox/pflash-request-probe
265k • Updated • 7
Lucebox/pflash-segment-probe
265k • Updated • 21
Lucebox/Qwen3.8-Flash-Next-GGUF
177B • Updated • 77 • 1
Lucebox/DeepSeek-V4.1-Flash-DSpark-GGUF
14B • Updated • 1.27k • 1
Lucebox/DeepSeek-V4.1-Flash-ROCMFP23-GGUF
755B • Updated • 42.1k • 8
Lucebox/Qwen3.8-27B-IQ4_XS-fast-GGUF
Image-Text-to-Text • 27B • Updated • 642 • 3
Lucebox/DeepSeek-V4-Flash-0731-ROCmFP3
284B • Updated • 6.09k • 25
Lucebox/Qwen3.8-27B-DFlash2-GGUF
2B • Updated • 508
Lucebox/DeepSeek-V4-Flash-0731-DSpark-GGUF
20B • Updated • 709 • 7
datasets 0
None public yet