| """Ask Jegue questions without the benchmark harness. |
| |
| Needs one CUDA GPU (about 20 GB), the Decision Index kit installed (it provides decision_index.engines.base) and the base model |
| Qwen/Qwen3.5-9B in the Hugging Face cache. From the repo root: |
| |
| PYTHONPATH=code python examples/ask.py |
| |
| The engine takes a shared context ("state") and a dict of questions. A "choice" question has option keys and descriptions and |
| returns one probability per option; a "noul" question is a yes/no and returns the probability of yes. |
| """ |
| import os |
| import sys |
|
|
| HERE = os.path.dirname(os.path.abspath(__file__)) |
| ROOT = os.path.dirname(HERE) |
| sys.path.insert(0, os.path.join(ROOT, "code")) |
|
|
| from decision_engine_dense import DenseSlotEngine |
|
|
| WEIGHTS = os.path.join(ROOT, "weights") |
| engine = DenseSlotEngine(model="Qwen/Qwen3.5-9B", run=WEIGHTS, ckpt=WEIGHTS) |
|
|
| state = "Email from a customer: 'My parcel was due on Monday and it is Friday. Tracking has not changed since Tuesday. I want my money back.'" |
| questions = { |
| "intent": { |
| "type": "choice", |
| "instructions": "What does the customer mainly want?", |
| "criteria": {"refund": "A refund", "status": "An update on the delivery", "complaint": "To complain about the service", "other": "Something else"}, |
| }, |
| "urgent": { |
| "type": "noul", |
| "instructions": "Should this email be escalated to a human agent today?", |
| }, |
| } |
|
|
| response, _ = engine(state, questions) |
| for qid, answer in response["answers"].items(): |
| if answer["type"] == "choice": |
| ranked = sorted(answer["probabilities"].items(), key=lambda kv: -kv[1]) |
| print(f"{qid}: {answer['choice']} " + " ".join(f"{k}={v:.2f}" for k, v in ranked)) |
| else: |
| print(f"{qid}: P(yes)={answer['noul']:.2f}") |
|
|
