Module vision 

Vision tower for Qwen3.8-27B multimodal input (lane/vision, 2026-08-15).

The qwen3_5_vision ViT (depth 27, hidden 1152, heads 16, gelu_pytorch_tanh, patch 16, spatial_merge 2, temporal_patch 2, LEARNED pos embeddings on a 48x48 grid) lives in the official checkpoint’s outside.safetensors (the unquantized shard) — the quantized trunks (ct-NVFP4 etc.) strip it. MEMRA_VISION_DIR points at any directory carrying that shard; the tower output is plain [n_tokens, out_width] embeddings (out_width = merger fc2 out_features, derived from the shard — 5120 on q38-27B, 2048 on Ornith-1.5-35B-A3B), so vision requests serve on any trunk whose n_embd matches the shard’s. Text side uses standard sequential rope (rope_scaling is null on these models — no M-RoPE), so spliced image tokens take ordinary positions.

v1 posture: correctness-first — cuBLASLt f32 GEMMs (Engine::linear + bias epilogue), sdpa_naive(causal=false) for the bidirectional attention, host-side permutes between stages (the tower is a small fraction of a vision request; optimize later). Parity gate: merger-output cosine vs the HF reference per VISION-LANE.md.

Structs

EmbedOverlay

Mixed-embedding prime overlay: image embeddings that replace <|image_pad|> token embeddings at prompt-relative positions during prime_cache_overlaid. rows holds all images’ merger outputs concatenated ([total_rows, n_embd]); each span is (prompt_pos, row_off, n_rows) — rows [row_off, row_off+n_rows) land at prompt positions [prompt_pos, prompt_pos+n_rows). Spans must not overlap.

VisionTower

Enums

OverlayPublish

Overlay publication policy (MEMRA_VISION_OVERLAY_PUBLISH, lane/glm53-vision-ppn): whether the tower’s rows are republished into the engine that owns embedding intake.

Constants

V_DEPTH

V_HEADS

V_HEAD_DIM

V_HIDDEN

V_INTER

V_MERGE

V_MERGED_IN

V_PATCH

V_PATCH_IN

V_POS_GRID

V_TEMPORAL

Functions

overlay_publications

Overlay publications this process has performed. A gate arm that means to exercise the publication path asserts this counter MOVED — a same-context run would otherwise take the zero-copy branch and print PASS having tested nothing (the non-vacuity law).

overlay_publish_mode

overlay_publish_resolve over the live environment.

overlay_publish_resolve

Pure resolver for MEMRA_VISION_OVERLAY_PUBLISH so unit tests pin the whole arm matrix without mutating process-global environment (the pp::peer_probe_startup_policy pattern). None = unset = Auto; auto/1 = Auto; force = Force; 0/off = Never. An unrecognized value is a REFUSAL, not a silent default — a mistyped door must never decide a correctness path quietly.