{ "schema_version": 1, "builder_plan": { "scenario": "A self-hosted assistant is too expensive because its model uses most of the GPU memory.", "change": "Try the paper's supported lower-precision weights on a copy of your serving setup.", "baseline": "The same model at its original precision, on the same GPU.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Peak memory, task pass rate, latency, and total serving cost at the same context and load.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2609.24757", "title": "NPU Accelerator: Quantized Real-Time Vehicle Detection on PYNQ-Z1 Using FINN", "summary": "The method replaces GPU- or CPU-based inference of full-precision detectors with a quantized (w2a4), QAT-trained lightweight YOLO variant compiled through Brevitas/QONNX and FINN onto a small Zynq FPGA dataflow accelerator. It costs substantial engineering effort across QAT, model slimming, FINN folding and FIFO sizing, and Vivado implementation, plus acceptance of 2-bit weights and modest accuracy (0.594 mAP@0.5). Failure modes include accuracy collapse from aggressive low-bit quantization on harder datasets, only one of many explored configurations meeting all constraints, toolchain fragility across Brevitas/FINN/Vivado versions, and results that do not transfer to other boards or input resolutions.", "rationale": "The approach requires dedicated FPGA hardware and a specialized co-design toolchain outside the reader's GPU/API infrastructure, so it is relevant to track but not adoptable or cheaply testable in their setting.", "technique": "Quantized FINN FPGA accelerator for real-time detection", "family": "quantizacao", "recommendation": "observar", "published": "2026-09-21", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.24757", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces GPU- or CPU-based inference of full-precision detectors with a quantized (w2a4), QAT-trained lightweight YOLO variant compiled through Brevitas/QONNX and FINN onto a small Zynq FPGA dataflow accelerator. It costs substantial engineering effort across QAT, model slimming, FINN folding and FIFO sizing, and Vivado implementation, plus acceptance of 2-bit weights and modest accuracy (0.594 mAP@0.5). Failure modes include accuracy collapse from aggressive low-bit quantization on harder datasets, only one of many explored configurations meeting all constraints, toolchain fragility across Brevitas/FINN/Vivado versions, and results that do not transfer to other boards or input resolutions.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.24757", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On a PYNQ-Z1 (Zynq XC7Z020), the LP-YOLO Slim configuration with 256x256 input, w2a4 quantization, and a 142.86 MHz PL clock achieves 35.66 FPS at 2.91 W (12.25 FPS/W), 45.11 ms PL latency, and 0.594 mAP@0.5 on Pascal VOC, the only tested configuration meeting all four deployment requirements simultaneously.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.24757", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }