Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
C
fix(family_registry): accept compress_ratios longer than num_hidden_layers
QQSHI1374637937fb5b
diff --git a/c/family_registry.py b/c/family_registry.pyindex f5ec4cfe..f46082b4 100644--- a/c/family_registry.py+++ b/c/family_registry.py@@ -364,17 +364,17 @@ def _dsv4_geometry(config, context, _model_dir):index_hd = _required_int(config, "index_head_dim", "deepseek_v4")ratios = config.get("compress_ratios")- if not isinstance(ratios, list) or len(ratios) != layers:+ if not isinstance(ratios, list) or len(ratios) < layers:raise ValueError(- "deepseek_v4: compress_ratios must be a list matching "- "num_hidden_layers")+ "deepseek_v4: compress_ratios must be a list of at least "+ "num_hidden_layers entries")# Fixed window ring: n_layers * window * head_dim fp32.fixed = layers * window * head_dim * 4# Compressor + indexer states, scaling with context / ratio.state = 0- for ratio in ratios:+ for ratio in ratios[:layers]:if not isinstance(ratio, bool) and isinstance(ratio, int) and ratio > 0:compressed = (context + ratio - 1) // ratiostate += compressed * head_dim * 4