Iteratively tune esp-ppq QuantizationSetting to recover post-quantization accuracy on ESP-DL targets. Drives a closed loop of "baseline -> calibration × TQT(default) cartesian product -> distribution-aware residual fixes -> agent-driven open exploration -> re-evaluate" in the current Python environment, using a minimal user contract (calib dataloader + evaluate function). Generic across architectures (ResNet / EfficientNet / ViT / DETR / YOLO / LSTM and any esp-ppq-supported graph) — the search procedure is distribution-driven and does not depend on a specific network family. Method ordering is accuracy-first with a soft penalty for passes that slow down on-device inference; once the prescribed Phase-1/2/3 sequence exhausts, the skill hands control to the agent (Phase 5) with a structured history of improving levers + the per-iteration error artifacts to read, so the agent can compose multi-knob iterations (lever stacking, calibration cross-pollination, ablation, cost-trim) without a rigid template. LSQ on POWER_OF_2 targets is auto-disabled (silent degenerate; use TQT instead) and esp32p4 layer-wise equalization is warn-only (esp-ppq officially "Not recommend" for per-channel weights but empirically can still help on some models). Use this skill whenever the user wants to improve a quantized esp-dl/.espdl model's accuracy, debug high quantization error, choose between calibration algorithms, decide on equalization / bias correction / weight split / mixed precision / TQT / blockwise reconstruction settings, or run an automated search over esp-ppq quantization options. Also triggers for "esp-ppq 量化调参", "降低量化误差", "Layerwise quantization error 分析", "QuantizationSettingFactory.espdl_setting 怎么调", "混合精度量化", "量化精度恢复", "espdl 量化优化".
☆1.09k repo stars⑂215 repo forks