Niko1221 0351ecb11d Two GTX 1080s are a working PC: an 8 GB card starts Off only beside a bigger card, and a PC whose only NVIDIA cards are Pascal uses them on the CUDA 12 engine (#3655) ...

recommend() (Terminal, app and shared vectors) now turns a card with 8 GB or less Off only when a bigger usable card is there or it has under 4 GB. The hardware report (tools/strata_mcp.py) applies setup's own opt-in rule for a PC with only Pascal/Volta NVIDIA cards and no usable AMD card, so those cards are no longer flagged 'older than the RTX 20 series' (engine_variant cuda12); the feed already picks the cuda12 package for sm_61 and refuses cuda13. Tests: shared gpu-roles vectors, strata_mcp hardware, feed selection for sm_61, and Tune keeping a two-card role map.

2026-10-22 13:47:41 +02:00

..

native_experts

Split GGUF artifacts: every shard found (a missing one is an error), tensors from their owning shard, per-role expert files (native_experts.txt v4)

2026-09-30 22:30:10 +02:00

gpu_roles_vectors.json

Two GTX 1080s are a working PC: an 8 GB card starts Off only beside a bigger card, and a PC whose only NVIDIA cards are Pascal uses them on the CUDA 12 engine (#3655)

2026-10-22 13:47:41 +02:00