From 3717083b4a4c383ad73f95eaa3dcf346ddf85c57 Mon Sep 17 00:00:00 2001 From: bong-water-water-bong Date: Sun, 15 Nov 2026 08:15:19 -0400 Subject: [PATCH] registry: answer the three NPU FLM packs the 11-15 census alerted on (engine#2528) Confucius4-R2T2 (Qwen3ASRForConditionalGeneration), Qwen3.8-27B-DFlash2 (DFlash2DraftModel) and LFM3-1.7B (Lfm3ForCausalLM) are mapped for hrx only, so census_watch prints them as NPU-route gaps. DFlash2 and LFM3 were quiet while unmapped (significant.json, engine#1451); once the registry mapped them, only registry/flm_answers.json can answer the gap, and it had no entry. Three entries plus their docs/npu.md rows; the 2026-11-05 table's registry column for the two now says they are mapped. Co-Authored-By: Claude Opus 5.5 --- docs/npu.md | 7 +++++-- registry/flm_answers.json | 13 +++++++++++++ 2 files changed, 18 insertions(+), 2 deletions(-) diff --git a/docs/npu.md b/docs/npu.md index c81673954..6f49e1a70 100644 --- a/docs/npu.md +++ b/docs/npu.md @@ -2396,8 +2396,8 @@ not a mapping: |---|---|---|---| | `FastFlowLM/MiniCPM-OCR-1.0-NPU2` | `MiniCPMOcrForConditionalGeneration` | none (the pin registers `MiniCPMForCausalLM`, `MiniCPM3` and `MiniCPMV4_6` only) | not served: an OCR/vision tower; the route claims text decoders only | | `FastFlowLM/MiniCPM-V-4.7-35B-A3B-NPU2`, `FastFlowLM/MiniCPM-V-4.7-1B-NPU2` | `MiniCPMV4_7ForConditionalGeneration` | none (V4_6's text tower maps through `qwen35`, a VLM wrapper the pin does not register) | not served: a vision-language pack; a converter registration for the text tower plus a vision scope decision come first | -| `FastFlowLM/Qwen3.8-27B-DFlash2-NPU2` | `DFlash2DraftModel` | none; reviewed not-an-alias (docs/arch-gaps.md, `dflash2`) | not served: a speculative drafter, not a standalone decoder; it needs the reviewed dflash2 support, not a mapping | -| `FastFlowLM/LFM3-1.7B-NPU2` | `Lfm3ForCausalLM` | none (LFM2 / LFM2-MoE are mapped; LFM3 is not registered at the pin) | not served: no LFM3 model type or kernels in the route | +| `FastFlowLM/Qwen3.8-27B-DFlash2-NPU2` | `DFlash2DraftModel` | none on 2026-11-05 (reviewed not-an-alias, docs/arch-gaps.md, `dflash2`); since mapped -> `dflash`, hrx only | not served: a speculative drafter, not a standalone decoder; it needs the reviewed dflash2 support, not a mapping | +| `FastFlowLM/LFM3-1.7B-NPU2` | `Lfm3ForCausalLM` | none on 2026-11-05 (LFM3 was not registered at the pin); since mapped -> `lfm3`, hrx only | not served: no LFM3 model type or kernels in the route | | `FastFlowLM/Hy-Memory-30B-A3B-NPU2` | none on 2026-11-05 (no config); `HYV3ForCausalLM` / model_type `hy_v3` by 2026-11-06 | `HYV3ForCausalLM` -> `hy_v3`, hrx only | not served: an hrx-only family; the 2026-11-06 dev-box run prints it as `!! NPU FLM ... npu route: no` | | `FastFlowLM/smolvla-NPU2` | none (no config), no `*.q4nx` | none | not a pack: a config-less VLA repo. It prints as `!! NEW FLM` because its config carries no `library_name` for the custom-code rule, but there is no architecture to map and no Q4NX file to serve | @@ -2454,6 +2454,9 @@ pack declares (a config with no `architectures` declares its `model_type`): | `Phi3ForCausalLM` | Phi4-mini-Instruct | not served: `phi3`, hrx only | engine#1786 | | `Gemma3TextModel` | Embedding-Gemma-300M | not served: `gemma-embedding`, an embedding model, hrx only | engine#1786 | | `WhisperForConditionalGeneration` | Whisper-V3-Turbo | not served: a speech encoder-decoder the registry does not map | engine#1783 | +| `Qwen3ASRForConditionalGeneration` | Confucius4-R2T2 (2026-11-15) | not served: a speech-recognition pack (audio encoder + text decoder), `qwen3vl`, hrx only | engine#1783 | +| `DFlash2DraftModel` | Qwen3.8-27B-DFlash2 | not served: a speculative drafter, `dflash`, hrx only (the first table above) | this section | +| `Lfm3ForCausalLM` | LFM3-1.7B | not served: `lfm3`, hrx only (the first table above) | this section | `FastFlowLM/smolvla-NPU2` is answered by its repo id (`repos` in the same file): it has no config and no `*.q4nx` file, so there is no architecture to key on and nothing for the route to load. diff --git a/registry/flm_answers.json b/registry/flm_answers.json index b1ca96217..8f236f4af 100644 --- a/registry/flm_answers.json +++ b/registry/flm_answers.json @@ -80,6 +80,19 @@ "answer": "not served: a speech encoder-decoder the registry does not map; the Whisper encoder on our runtime is engine#1783", "doc": "docs/npu.md#fastflowlms-npu2-drops-the-families-the-route-does-not-serve", "issue": 1783 + }, + "Qwen3ASRForConditionalGeneration": { + "answer": "not served: a speech-recognition pack (audio encoder + text decoder); `qwen3vl`, hrx only, and the route claims text decoders only", + "doc": "docs/npu.md#fastflowlms-npu2-drops-the-families-the-route-does-not-serve", + "issue": 1783 + }, + "DFlash2DraftModel": { + "answer": "not served: a speculative drafter, not a standalone decoder; `dflash`, hrx only", + "doc": "docs/npu.md#fastflowlms-npu2-drops-the-families-the-route-does-not-serve" + }, + "Lfm3ForCausalLM": { + "answer": "not served: `lfm3`, hrx only; no LFM3 model type or kernels in the route", + "doc": "docs/npu.md#fastflowlms-npu2-drops-the-families-the-route-does-not-serve" } }, "repos": {