pyr0ball 13ca082a43 chore(models): refresh model registries with current cluster catalog ... Replace stale llama/mistral/phi model refs with models active on the cluster: deepseek-r1 (1.5b, 7b-4bit, 0528-qwen3-8b-gguf), granite-4.1-8b, qwen2.5 (3b, 7b), capybarahermes-2.5-mistral-7b, darwin-9b-opus. Update benchmark_plans.py doc examples to match.		2026-05-17 11:24:03 -07:00
..
__init__.py	feat: initial avocet repo — email classifier training tool	2026-02-27 14:07:38 -08:00
benchmark_classifier.py	feat(benchmark): wire EmbeddingKNNAdapter into MODEL_REGISTRY as embed-knn-nomic	2026-05-05 12:43:48 -07:00
benchmark_plans.py	chore(models): refresh model registries with current cluster catalog	2026-05-17 11:24:03 -07:00
benchmark_style.py	refactor(bench): extract benchmark tabs — classifier, compare, llm-eval, style, voice	2026-04-24 14:56:17 -07:00
benchmark_voice.py	refactor(bench): extract benchmark tabs — classifier, compare, llm-eval, style, voice	2026-04-24 14:56:17 -07:00
classifier_adapters.py	fix(classifier): majority-vote key, partial-load guard, sparse label test	2026-05-05 11:39:24 -07:00
export_plans.py	feat: plans benchmark harness — model scoring for CF planning prompts	2026-05-02 23:36:04 -07:00
finetune_classifier.py	fix(avocet): use_reentrant=False for gradient checkpointing	2026-03-15 17:23:40 -07:00
gather_corpus.py	chore: add README + gather_corpus.py script	2026-04-24 15:29:26 -07:00
sft_import.py	fix: log on malformed json in _read_jsonl, use streaming id dedup	2026-04-08 07:37:22 -07:00