Local LLM inference on one workstation. Measurements in rig-log.
Merged upstream
- ik_llama.cpp: DeepSeek-V4.1 support (#2455); fixes #2513, #2511, #2508, #2501, #2493, #2444, #2443, #2436
- llama.cpp: #29008
- cuda-oxide: #1314
- exllamav3: #376
- wails: #6000, #6006
Projects




