Emit numeric per-task scores in inference tasks map

Return score-only baseline and grader payloads

Output score-only task metrics in inference results

Harden API score fields to strict open interval

Clamp grader component scores to open interval

Fix baseline loop indentation in server app

Clamp emitted per-step rewards to strict open interval

Avoid boundary totals in fallback grading paths

Use 1e-3 epsilon for all submission score clamps

Clamp baseline endpoint scores to strict open interval

Clamp all server task scores to strict open interval

Clamp inference output scores to strict open interval

Force proxy-visible LLM probe in inference

Fix server app indentation error

Enforce strict (0,1) task scores

Probe the LLM proxy at startup

Use injected API_KEY for baseline LLM calls

fix reset endpoint indentation and empty-body compatibility

accept empty POST body on reset endpoint

hotfix: make space public and speed docker build for submission checks

align inference env vars with submission checklist

merge hfspace history and sync root runtime/docs with latest fixes

rewrite README with full end-to-end OSS docs and HF deployment setup

fix space UI routing and add structured inference markers

Add HF Space README metadata

Sync Space root with clinical-trial-triage

Sync Space with latest clinical-trial-triage

Add inference-backed UI mode and endpoint; clean triage docs

uii

Fix Space short_description length

Sync full clinical-trial-triage project into Space

Fix HF Space metadata fields

Add FastAPI Docker app for Space

initial commit

Update README.md

updated requirements based files

Revise README for Clinical Trial Triage project

Make inference fail-safe and deterministic; align README

started the project

Initial commit