Baekpica/dots3-note-prev-Mixed-Quant-GGUF

Report the 2026-09-06 ds4-dfm-rs optimization rounds

Rename GGUF architecture to dots3note

Note this GGUF is ds4-native, not llama.cpp

Add GitHub Sponsors button

Add support link

Document verified 512K ds4 serving

docs: publish verified ds4 Spark throughput

Document native ds4 support

Upload mixed-quant shard 10/10

Upload mixed-quant shard 9/10

Upload mixed-quant shard 8/10

Upload mixed-quant shard 7/10

Upload mixed-quant shard 6/10

Upload mixed-quant shard 5/10

Upload mixed-quant shard 4/10

Upload mixed-quant shard 3/10

Upload mixed-quant shard 2/10

Upload mixed-quant shard 1/10

Upload MQ87-VERIFY.json

Upload MQ87-SHA256SUMS

Record exact mixed-quant byte counts

Track GGUF with Git LFS

Add Apache-2.0 license

Add mixed-quant model card

initial commit