decider: reuse request prefixes with conservative auto selection

README.md

@@ -169,6 +169,7 @@ pinned runtime for text, image and video state:
 | Model | Status |
 | --- | --- |
 | LAYA | [External worker](recipe/laya/README.md); [Python worker on Apple Silicon (MPS) and CPU](recipe/laya/apple-silicon.md); [CPU checkpoint reader](src/models/laya/README.md); [native Rust/CUDA worker on Hopper](recipe/laya/native/README.md) |
+| Decider-2B v11 | [Native Rust/CUDA text worker](recipe/decider/README.md); Choice/Noul/isolated Score; opt-in batching, Graph and request-local prefixes; [validation scope](recipe/decider/validation.md) |
 | Cua-S1 4B 0.2 (`text` adapter) | [Python worker](recipe/cua_s1/text.md); [native worker](recipe/cua_s1/native.md), CUDA; unmerged profile validated on the cumulative A800 sm_80 build, earlier merged path on sm_89 |
 | Cua-S1 4B 0.2 (`multimodal` adapter) | [Python CUDA reference](src/frontend/cua_s1.py) and [native Rust/CUDA screenshot worker](recipe/cua_s1/native_screenshot.md); one PNG/JPEG image, `choice`; scoped RTX 4090 validation |
 | JEMM (Qwen3.8-27B) | [Native Rust/CUDA worker](recipe/jemm/README.md); [text and 1–4-image frozen-corpus parity on A800](src/models/jemm/README.md#validation) |

docs/architecture.md

@@ -7,15 +7,16 @@ design. Concrete input/output types follow each executor's supported layout.
 ## Implementation status
 
 The [Rust frontend](../src/frontend/README.md) currently forwards HTTP requests
-to separately running workers. Cua-S1 and Open-Jev have native Rust/CUDA workers
+to separately running workers. Cua-S1, Open-Jev and Decider have native Rust/CUDA workers
 that share the [Qwen3.5/3.8 executor](../src/models/qwen3_5/native/), which accepts
 single prompts and bounded packed prefill. Cua-S1 and Open-Jev-9B use
 single-prompt calls; Open-Jev-27B-v1.1 packs candidates within one request for
 input and gate/up GEMMs while preserving per-sequence mixers and output/down GEMM
-shapes. The executor can also run prompts that share token prefixes, each prefix
-once (below); no worker uses that yet.
+shapes. Decider optionally packs up to four complete rows and 4096 tokens within
+one request, or selects request-local shared prefixes through the same canonical
+continuation path used by JEV-VL (below). Both optimizations default off.
 [Laya's native worker](../src/models/laya/README.md)
-uses a separate Hopper CUDA backend for one complete padded request. All three
+uses a separate Hopper CUDA backend for one complete padded request. These workers
 coordinate independent processors and executors through
 `prepare` → `execute` → `finish` and use the
 [native runtime](../src/runtime/README.md) for FIFO admission and blocking dispatch
@@ -28,7 +29,9 @@ FP32 adapters and framework-reference arithmetic. It computes selected BF16
 label-head logits on CUDA, then calibrates typed answers on the CPU. JEMM admits
 one complete text or 1–4-image request and synchronizes both model streams before
 releasing its resources. See its [model contract](../src/models/jemm/README.md).
-Laya computes its scorer and action head on CUDA. Its fixed-shape
+Decider projects selected tied-embedding rows on CUDA with BF16 output rounding
+before FP32 calibration and CPU response assembly. Laya computes its scorer and
+action head on CUDA. Its fixed-shape
 Graph captures Encoder/Decision; gather, scorer/action head and synchronized
 readback remain outside capture. Native Metal remains planned; Python workers
 retain their documented reference/serving roles.
@@ -63,6 +66,7 @@ and candidate identity, usage, and response metadata outside the executor.
 | --- | --- | --- | --- |
 | Cua-S1 | One unpadded token-ID vector and option count per question, in request order. | One FP32 answer-letter logit vector per question. | Per-question softmax, choice/confidence, ordered answers, and token usage. |
 | Open-Jev | Token-ID vectors grouped by question, then independent candidate, in request order. | One FP32 learned scalar per candidate in the same grouping. | Add the `noul` false logit of zero, calibrate across each complete question, and restore typed answers, usage, and metadata. |
+| Decider | Complete independent token-ID rows in question order; Score expands one no/yes row per level. | One BF16-projection-rounded FP32 selected-label logit vector per row. | Per-type calibration, whole-Score normalization, ordered typed answers and unique-prefix token usage. |
 | Laya | One padded request: token IDs, true lengths, question types and ordered option markers; at most 16 questions, 512 tokens per row and 2048 markers. | Per-question FP32 option logits and two action logits copied back after GPU heads. | Calibrate and decode ordered `choice`, `score` and `noul` answers, usage and metadata. |
 | JEMM | Ordered token IDs, three-axis positions, candidate counts and optional shared image patches for one complete request. | Per-question selected BF16 label logits copied back as FP32. | Apply checksum-covered text/image calibration and reconstruct ordered `choice`, `score` and `noul` answers and usage. |
 
@@ -72,9 +76,11 @@ Open-Jev's model-specific batch adapter packs up to 16 independent candidates an
 execute alone. It restores question/candidate grouping before normalization.
 Laya batches questions within one request. Shared runtime
 admission precedes blocking dispatch: Cua-S1 admits one question forward at a
-time; Open-Jev and Laya admit one complete request. Cua-S1's CPU letter projection
+time; Open-Jev, Decider and Laya admit one complete request. Cua-S1's CPU letter projection
 stays outside admission; Open-Jev's scalar heads and Laya's GPU heads and
-synchronized readback remain inside their request unit. Qwen model mutexes guard
+synchronized readback remain inside their request unit. Decider includes every
+backbone row and CUDA head projection, synchronizes both streams and retires
+failed state before releasing admission. Qwen model mutexes guard
 mutable state. Laya's dedicated owning thread confines its non-Send CUDA state
 and receives admitted work over a rendezvous channel. Cancellation after
 dispatch retains the scheduler permit until execution completes. Laya

docs/supported-models.md

@@ -19,6 +19,7 @@ Models that are being added are also tracked in issues labeled [new model](https
 | Valen-Preview-0923 | [Python reference worker](../recipe/valen/reference.md); text state or one inline PNG/JPEG image; `choice` questions | Not supported | Author-run on RTX 5060 Laptop (sm_120, WSL2): one text and one 2×2 PNG request; input and readout parity ≤1e-6 against reference (model outputs not compared). Owner: community. | Not supported | Python 3.12, pinned Valen source/checkpoint, Qwen3.5-2B base, BF16 CUDA |
 | Open-Jev-27B-v1.1 | [Native Rust/CUDA worker](../recipe/open_jev/native.md) on the shared Qwen3.5/3.8 executor | Not supported | Validated on H200 (sm_90) for the [74 single-candidate workload](../recipe/open_jev/validation.md) | Not supported | Compute capability 8.0 or newer, CUDA toolkit to build, exported merged weights and trained head |
 | Open-Jev-9B | The same [native Rust/CUDA worker](../recipe/open_jev/native.md) | Not supported | Validated on compute capability 8.9 against the reference for [253 requests](../recipe/open_jev/validation-9b.md) | Not supported | Compute capability 8.0 or newer, CUDA toolkit to build, exported merged weights and trained head |
+| Decider-2B v11 | [Native Rust/CUDA worker](../recipe/decider/README.md) | Contract checks only; no CPU inference | RTX 4090 (`sm_89`) synthetic parity corpus; [limits and provenance](../recipe/decider/validation.md) | Not supported | Pinned original BF16 checkpoint; ABI 9 library and consumers rebuilt together |
 | JEMM / Qwen3.8-27B | [Native Rust/CUDA worker](../recipe/jemm/README.md) | Host checks only; no CPU inference | A800 (`sm_80`) [text and 1–4-image frozen corpus](../src/models/jemm/README.md#validation) | Not supported | Pinned `jemm-native/2` export with BF16 base and unmerged FP32 LoRA, selected untied LM head, cuDNN 9 and additive reference CUDA symbols |
 | autotrust/JEV-27B-VL | [Experimental native worker](../recipe/jev_vl/README.md), single-question API | Contract tests only; no CPU inference | H800 frozen-corpus check; [experimental scope and limits](../recipe/jev_vl/validation.md) | Not supported | Pinned merged checkpoint, rebuilt ABI 9 CUDA library; image inputs require offline Transformers preencoding |
 | CLM-v0.1-8B | [External `clm-serve` recipe](../recipe/clm/README.md) with a CPU stub embeddings server | **Stub-encoder contract checks only** ([#23](https://github.com/ThinkFlowLab/system1-omni/pull/23)); not real Qwen3-8B decisions | Real encoder unverified by the merged recipe | Unverified | Python, upstream CLM and head checkpoint; a real encoder requires a separate embeddings server |

mkdocs.yml

@@ -72,6 +72,8 @@ nav:
       - Open-Jev native text worker: recipe/open_jev/native.md
       - Open-Jev-27B H200 validation: recipe/open_jev/validation.md
       - Open-Jev-9B validation: recipe/open_jev/validation-9b.md
+      - Decider native text worker: recipe/decider/README.md
+      - Decider validation: recipe/decider/validation.md
       - JEMM native worker: recipe/jemm/README.md
       - JEV-27B-VL experimental worker: recipe/jev_vl/README.md
       - JEV-27B-VL validation status: recipe/jev_vl/validation.md

recipe/README.md

@@ -22,6 +22,8 @@ For a first real decision, follow the [complete CPU walkthrough](../docs/getting
   backbone and trained decision head of Open-Jev-27B-v1.1 or Open-Jev-9B, then
   serve with Rust and CUDA. [Open-Jev-9B validation](open_jev/validation-9b.md)
   compares the 9B worker with the reference.
+- [Decider-2B v11 native worker](decider/README.md): pinned BF16 text decisions,
+  optional request-local batching, Graph and prefix reuse; [validation](decider/validation.md).
 - [JEMM native worker](jemm/README.md): pinned text/image decisions and export;
   known multi-image numerical limitation is documented in the model contract.
 - [JEV-27B-VL experimental worker](jev_vl/README.md): native text and preencoded