fix(prithvi): drop .view() on lists; enrich warmup traceback

fix(prithvi): dict-aware normalizer; expose riprap-models last_errors

fix(eo): unblock TerraMind LoRA + Prithvi v2 inference on the L4 Space

fix: thread Mellea attempt index + diagnose riprap-models 500s

fix(synthesis): handle 3-D DEM from _read_dem_patch (it returns (1,H,W))

feat: every EO specialist's prediction renders as a map layer

fix(synthesis): DEM shape is (1, 1, H, W) per encoder probe

fix: synthesis DEM shape matches local 3-D contract

feat: terramind_synthesis now routes through droplet remote inference

fix(droplet): _build_chip_tensor handle 5D input from eo_chip_cache

sync: bring HF Space to parity with origin/main (squash, XET-backed)

deploy: sync all changes from main at 6904684

Voice pass: strip em-dashes from user-facing docs

Self-contained droplet redeploy: Dockerfile + bring-up script

feat: route all GPU-accelerable inference through MI300X (Phase 1+2 of full GPU)