| | Card 2026-09-07 (evening): two-box pipelined prefill in the served path; prefill defaults landed; DFlash2 p512/gen64 diagnostic corrected to 59.5 tok/s | avlp12 | Sep 7, 2026 |
| | Card 2026-09-07: multi-turn session reuse row (99.3% cached, 0.17 s TTFT), branch head a6634a75; earlier sections unchanged | avlp12 |
| | Refresh chat_template.jinja to upstream 690b7052; card note on what changed (image markers, clear_thinking default, tool-result guards) | avlp12 |
| | Update model card with serving measurements and cache caveats | avlp12 |
| | recipe table: KDA attention projections/gates are 8-bit (DSA q_a/q_b/o 6-bit; kv_a/indexer 8-bit) -- corrected from a full safetensors header scan | avlp12 |
| | prefill@65k 314 tok/s: 2^31 query-chunk boundary fix + gate retune; buffer env superseded by fused kernels | avlp12 |
| | buffer env caveat: it is a B=1 lever, regresses batched serving (-7.5% at B=16) | avlp12 |
| | gather gate: PR 2087 now defaults to 32768 upstream (adopted); note config-dependence | avlp12 |
| | card: performance chart (first upload vs now) | avlp12 |
| | card: pipeline_tag text-generation (HF MLX LM app filter requires it; vision usage documented in body) | avlp12 |