anvesa-code-400m

A code-search encoder for anvesa: Ettin-encoder-400m (jhu-clsp, MIT) fine-tuned on 105 MIT-licensed repositories, learning from the 1B fine-tune as a teacher. ONNX, fp32: model.onnx with its weights in model.onnx_data, and tokenizer.json. Mean pooling over real tokens, L2-normalised, 512 tokens, no prefix. It embeds anvesa's symbol cards without their WQL path line.

ranker-noword.json is the XGBoost ranker fit on this encoder's scores for anvesa's experimental.learnedRanker (20 features, no word overlap).

	corpus test (1,710 generated queries, 57 repos, 13 languages)	bench (241 hand-written queries, 8 repos)
first stage, hub penalty 0.3	0.349	0.684
with the ranker	0.369	0.682

CPU: 98 ms a card (ONNX Runtime, 12 threads). For comparison, all-MiniLM-L6-v2 scores 0.293 and 0.547 without a ranker.

anvesa model install anvesa-code-400m --download