fix: remap bare Sequential keys to net.* when loading ddg_da.pt

fix: patch torchmetrics._apply using CPU probe tensor instead of self.device

fix: use callable map_location to skip Lightning model.to() — avoids torchmetrics CUDA assertion

cleanup: remove debug tracing, restore clean health endpoint

fix: patch raw checkpoint to replace all CUDA refs before load

fix: patch torchmetrics._apply for CPU-only load of GPU-trained checkpoint

debug: global exception handler returns JSON traceback

debug: expose full traceback on predict errors

debug: attempt model download in /health and surface errors

debug: add model status to /health endpoint

fix: download models from HF Hub before existence check

fix: track app/models/ package, ignore only top-level /models/

fix: move Optional import to top of feature_engineer.py

Add Space README config

StockPro ML backend with pytorch-forecasting TFT