Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
Python calibrationclassificationdecision-modelhuggingfacejevmodernbertmultilingualnlppythonpytorchroutingtyped-decisionszero-shot
Commits on main
fix(serve): a lone surrogate in the body was a 500 instead of a 400
06462bf PerryLink committed 1 month ago
docs(reference): order the page so the "below" reference is true
8c965bf PerryLink committed 1 month ago
research: refresh the 51-language multilingual columns from the re-run (#208)
ad28be4 PerryLink committed 1 month ago
fix(ts): stringify score legend values and document serve choice limit (#302)
655a482 Swatantra Yadav committed 1 month ago
fix(router): keep the per-call hooks out of the docstring's Args list
d0339e9 Aashish committed 1 month ago
docs(readme): point the long-document section at router.predict_long
d6a7d5f Aashish committed 1 month ago
fix(agent): a choice criteria list that cannot give every option its own answer key
cc622bf Aashish committed 1 month ago
feat(router): predict_long -- route, then scan the whole state in windows
9810daa Aashish committed 1 month ago
feat(onnx): add ONNXAgent.predict_long, the long-document API the torch Agent has
3217741 Aashish committed 1 month ago
ci: use per-commit concurrency group for pushes to main (#399)
b750d17 Raj Ojha committed 1 month ago
fix(ts): accept pydantic v2 Optional[...] (anyOf/oneOf nullable) in structured schemas
4942728 Raj Ojha committed 1 month ago
feat(ts): port mixed-segment detection from Python analyse() (#384)
e37428a Swatantra Yadav committed 1 month ago
fix(mcp): run agent in auto mode when no router is loaded
09ab974 soumojit-D48 committed 1 month ago
feat(onnx): add ONNXAgent.predict_batch, the batch API the torch Agent has (#323)
4b81bd7 Aashish committed 1 month ago
feat(serve): report per-agent CPU fallback counters in /health
7c923ea umran666 committed 1 month ago
review: narrow lang fallback, docs, and flag before end hooks (#361)
df5d655 Aashish committed 1 month ago
fix(lint): import confidence helpers directly and remove unused symbols
76bfcf0 Aashish committed 1 month ago
feat: confidence-based abstention gating on answer_confidence (min_confidence) (#361)
0c6fc6a Aashish committed 1 month ago
fix(mcp): run Agent directly for model=auto and report null model
40f397d umran666 committed 1 month ago
fix(ts): resolve example weights relative to the script, fall back to HF
dc026de nnlgsakib committed 1 month ago
fix(research): don't echo exception text in the capacity bench server's 400
61f42e8 Bhushan Kinge committed 1 month ago
docs(benchmarks): keep a non-laya SDK out of the capacity environment record
93397be Bhushan Kinge committed 1 month ago
docs(benchmarks): back the NVIDIA capacity section with a committed results file and script
721cf9f Bhushan Kinge committed 1 month ago