{ "source": "jevision-engine-v2-full-eval-20261006.json", "source_sha256": "c648a0c257b30fe8130511703b458febf7dc009f9c320ad5a474b34411605b44", "hardware": "RTX 5060 Ti 16 GB", "latency": { "count": 100, "median_ms": 355.9841180103831, "mean_ms": 279.35151436482556, "p90_ms": 401.043732999824, "p95_ms": 407.9685560427606, "max_ms": 426.7997309798375, "within_400ms": 89, "within_500ms": 100, "within_600ms": 100, "within_700ms": 100 }, "request_count": 100, "judgment_count": 1563, "decision_agreement": 0.9872040946896993, "labeled": { "states": 900, "questions": 2423, "accuracy": { "all": { "original": 0.9413949649195212, "engine": 0.9413949649195212, "ratio": 1.0 }, "domain:sketch": { "original": 0.8714090287277702, "engine": 0.8727770177838577, "ratio": 1.001569858712716 }, "labels:teacher-silver": { "original": 0.9059689288634505, "engine": 0.9067865903515945, "ratio": 1.0009025270758123 }, "domain:photo": { "original": 0.9393939393939394, "engine": 0.9381313131313131, "ratio": 0.9986559139784946 }, "labels:vqa-human-consensus": { "original": 0.91, "engine": 0.9066666666666666, "ratio": 0.9963369963369962 }, "domain:text": { "original": 1.0, "engine": 1.0, "ratio": 1.0 }, "labels:deterministic": { "original": 1.0, "engine": 1.0, "ratio": 1.0 } }, "nll": { "original": 0.21043879280027272, "engine": 0.20826901852354016 }, "brier": { "original": 0.0944555328427415, "engine": 0.09420396303044455 }, "changed_choices": 4, "max_probability_delta": 0.15646521747112274, "mean_probability_delta": 0.0005122810770939374, "state_latency": { "count": 900, "median_ms": 75.74560452485457, "mean_ms": 64.05471320483937, "p90_ms": 88.83039897773415, "p95_ms": 91.0153309814632, "max_ms": 98.99667103309184, "within_400ms": 900, "within_500ms": 900, "within_600ms": 900, "within_700ms": 900 } } }