Halve aca-model production assets batch size to fit V100 16GB

Reduce n_assets_batch_size from 2 to 1 in MODEL_CONFIG so the assets
state axis is streamed one slice at a time, lowering peak GPU memory
during solve on the V100-PCIE-16GB. Benchmark grid config is unchanged.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: hmgaudecker

src/aca_model/config.py

@@ -36,7 +36,7 @@ class GridConfig:
     # `batch_size` on the assets grid: chunked vmap stride for the
     # outer state loop. Useful at prod sizes for memory reasons; set
     # to 0 in BENCHMARK_GRID_CONFIG to skip the Python-loop overhead.
-    n_assets_batch_size: int = 2
+    n_assets_batch_size: int = 1
 
 
 MODEL_CONFIG = ModelConfig()