Qyvos v1: Julia-1 backbone (bit-exact) + Open-Jev head fine-tune (30k rows, low-RAM protocol)
31f7037 verified | """Named typed questions over the existing Julia scoring API.""" | |
| import math | |
| from .data import validate_row | |
| def predict_typed(engine, state, questions): | |
| if not isinstance(questions,dict) or not questions: | |
| raise ValueError('questions must be a nonempty mapping') | |
| rows=[];metadata=[] | |
| for qid,q in questions.items(): | |
| if not isinstance(qid,str) or not qid or not isinstance(q,dict): | |
| raise ValueError('Questions require nonempty string IDs and question objects') | |
| kind=q.get('type');criteria=q.get('criteria') | |
| if kind=='choice': | |
| if not isinstance(criteria,dict) or any(not isinstance(k,str) or not k for k in criteria): | |
| raise ValueError('Choice criteria must map nonempty IDs to descriptions') | |
| keys=list(criteria);labels=list(criteria.values()) | |
| elif kind=='score': | |
| if not isinstance(criteria,list):raise ValueError('Score requires an ordered rubric') | |
| labels=criteria;keys=[str(i) for i in range(len(labels))] | |
| elif kind=='noul': | |
| keys=['false','true'] | |
| if criteria is None: | |
| labels=list(keys) | |
| else: | |
| if not isinstance(criteria,dict) or set(criteria)!=set(keys): | |
| raise ValueError('Noul criteria must map false and true to descriptions') | |
| labels=[criteria[key] for key in keys] | |
| else:raise ValueError('Unsupported question type') | |
| row=dict(state=state,question=q.get('instructions'),type=kind,options=labels) | |
| validate_row(row,len(rows)+1) | |
| rows.append(row);metadata.append((qid,kind,keys)) | |
| scores=engine.logits(rows) | |
| if len(scores)!=len(rows):raise ValueError('Model returned an incorrect answer count') | |
| answers={} | |
| for (qid,kind,keys),z in zip(metadata,scores): | |
| if len(z)!=len(keys) or not all(math.isfinite(x) for x in z):raise ValueError('Invalid model scores') | |
| maximum=max(z);p=[math.exp(x-maximum) for x in z];total=sum(p);p=[x/total for x in p] | |
| result=dict(type=kind,probabilities=dict(zip(keys,p))) | |
| if kind=='choice':result['choice']=keys[max(range(len(p)),key=p.__getitem__)] | |
| elif kind=='score':result['score']=sum(i*x for i,x in enumerate(p)) | |
| else:result['noul']=p[1] | |
| if kind!='noul':result['max_probability']=max(p) | |
| answers[qid]=result | |
| return dict(answers=answers) | |