{ "n": 200, "images": 73, "accuracy_all": 0.81, "accuracy": { "noul": 0.8202247191011236, "choice": 0.8018018018018018 }, "nll": 0.4400090990936913, "brier": 0.25543111423553433, "ece": 0.06582706868648526, "ece_bins": 15, "reliability": [ { "bin": [ 0.2, 0.3 ], "n": 1, "confidence": 0.28821372985839844, "accuracy": 1.0 }, { "bin": [ 0.3, 0.4 ], "n": 5, "confidence": 0.3615827262401581, "accuracy": 0.4 }, { "bin": [ 0.4, 0.5 ], "n": 7, "confidence": 0.4423674430165972, "accuracy": 0.42857142857142855 }, { "bin": [ 0.5, 0.6 ], "n": 33, "confidence": 0.5360189441478613, "accuracy": 0.6363636363636364 }, { "bin": [ 0.6, 0.7 ], "n": 26, "confidence": 0.6526116694395359, "accuracy": 0.6923076923076923 }, { "bin": [ 0.7, 0.8 ], "n": 15, "confidence": 0.7433490335941315, "accuracy": 0.8 }, { "bin": [ 0.8, 0.9 ], "n": 38, "confidence": 0.8569445978654059, "accuracy": 0.8421052631578947 }, { "bin": [ 0.9, 1.0 ], "n": 75, "confidence": 0.953959968884786, "accuracy": 0.9733333333333334 } ], "thresholds": { "0.6": { "coverage": 0.77, "accuracy": 0.8766233766233766 }, "0.7": { "coverage": 0.64, "accuracy": 0.9140625 }, "0.8": { "coverage": 0.565, "accuracy": 0.9292035398230089 }, "0.9": { "coverage": 0.375, "accuracy": 0.9733333333333334 } }, "by_source": { "ai2d": { "n": 2, "images": 1, "accuracy": 0.5, "nll": 1.2898404388867466, "ece": 0.3732716739177704 }, "aokvqa": { "n": 4, "images": 2, "accuracy": 0.75, "nll": 0.611274152606737, "ece": 0.2411895990371704 }, "coco-exist": { "n": 39, "images": 25, "accuracy": 0.9230769230769231, "nll": 0.23563569869662415, "ece": 0.0540550612868407 }, "coco-position": { "n": 26, "images": 18, "accuracy": 0.9230769230769231, "nll": 0.20050254680859508, "ece": 0.12922574006594145 }, "coco-relation": { "n": 10, "images": 5, "accuracy": 0.8, "nll": 0.4535431995357432, "ece": 0.29919506013393404 }, "coco-size": { "n": 2, "images": 2, "accuracy": 1.0, "nll": 0.3333616534201184, "ece": 0.2508668899536133 }, "gqa-v2": { "n": 25, "images": 7, "accuracy": 0.68, "nll": 0.5026796189257893, "ece": 0.21928748846054075 }, "tqa": { "n": 10, "images": 2, "accuracy": 0.5, "nll": 1.2075195747510135, "ece": 0.35332188010215765 }, "vqav2-choice": { "n": 35, "images": 21, "accuracy": 0.8857142857142857, "nll": 0.3503283097869317, "ece": 0.15751930900982447 }, "vqav2-yesno": { "n": 38, "images": 24, "accuracy": 0.7368421052631579, "nll": 0.5690740238980908, "ece": 0.11526736793549436 }, "vsr": { "n": 9, "images": 4, "accuracy": 0.7777777777777778, "nll": 0.5381709545359337, "ece": 0.2373169246647093 } }, "by_kind": { "ai2d": { "n": 2, "images": 1, "accuracy": 0.5, "nll": 1.2898404388867466, "ece": 0.3732716739177704 }, "animal": { "n": 2, "images": 2, "accuracy": 1.0, "nll": 0.08400760063551806, "ece": 0.07780307531356812 }, "aokvqa": { "n": 4, "images": 2, "accuracy": 0.75, "nll": 0.611274152606737, "ece": 0.2411895990371704 }, "color": { "n": 14, "images": 11, "accuracy": 0.8571428571428571, "nll": 0.3618174527077554, "ece": 0.09759957662650516 }, "count": { "n": 10, "images": 7, "accuracy": 0.8, "nll": 0.6097179990869517, "ece": 0.3152927696704865 }, "exist": { "n": 39, "images": 25, "accuracy": 0.9230769230769231, "nll": 0.23563569869662415, "ece": 0.0540550612868407 }, "food": { "n": 1, "images": 1, "accuracy": 1.0, "nll": 0.0042057737329474765, "ece": 0.00419694185256958 }, "gqa": { "n": 11, "images": 6, "accuracy": 0.7272727272727273, "nll": 0.5031112251493984, "ece": 0.2824250134554776 }, "material": { "n": 2, "images": 2, "accuracy": 1.0, "nll": 0.22872908175288753, "ece": 0.19662755727767944 }, "meat": { "n": 1, "images": 1, "accuracy": 1.0, "nll": 0.1798803908020786, "ece": 0.16462987661361694 }, "position": { "n": 26, "images": 18, "accuracy": 0.9230769230769231, "nll": 0.20050254680859508, "ece": 0.12922574006594145 }, "relation": { "n": 10, "images": 5, "accuracy": 0.8, "nll": 0.4535431995357432, "ece": 0.29919506013393404 }, "room": { "n": 1, "images": 1, "accuracy": 1.0, "nll": 0.005263986423792628, "ece": 0.005250155925750732 }, "size": { "n": 2, "images": 2, "accuracy": 1.0, "nll": 0.3333616534201184, "ece": 0.2508668899536133 }, "spatial": { "n": 14, "images": 5, "accuracy": 0.6428571428571429, "nll": 0.5023404997500963, "ece": 0.1696794331073761 }, "tqa": { "n": 10, "images": 2, "accuracy": 0.5, "nll": 1.2075195747510135, "ece": 0.35332188010215765 }, "vegetable": { "n": 4, "images": 4, "accuracy": 1.0, "nll": 0.07101074950722219, "ece": 0.06772895157337189 }, "vsr": { "n": 9, "images": 4, "accuracy": 0.7777777777777778, "nll": 0.5381709545359337, "ece": 0.2373169246647093 }, "yesno": { "n": 38, "images": 24, "accuracy": 0.7368421052631579, "nll": 0.5690740238980908, "ece": 0.11526736793549436 } }, "by_axis": { "-": { "n": 162, "images": 67, "accuracy": 0.7901234567901234, "nll": 0.47554891340501837, "ece": 0.05595421993438109 }, "lr": { "n": 27, "images": 18, "accuracy": 0.8888888888888888, "nll": 0.3277449296076739, "ece": 0.19794687407988087 }, "tb": { "n": 11, "images": 11, "accuracy": 0.9090909090909091, "nll": 0.1921620679743721, "ece": 0.09886410019614479 } }, "by_options": { "2": { "n": 157, "images": 64, "accuracy": 0.8280254777070064, "nll": 0.38058168910086837, "ece": 0.06867314277181197 }, "3": { "n": 10, "images": 9, "accuracy": 0.9, "nll": 0.4393064845098579, "ece": 0.22365574836730956 }, "4": { "n": 31, "images": 18, "accuracy": 0.6774193548387096, "nll": 0.7616504939115978, "ece": 0.10811415699220474 }, "5": { "n": 2, "images": 2, "accuracy": 1.0, "nll": 0.12313223677190567, "ece": 0.1151973307132721 } }, "noul_confusion": { "tp": 42, "fp": 10, "fn": 6, "tn": 31, "yes_ratio": 0.5842696629213483 }, "request_latency_ms": { "p50": 174.91745902225375, "p95": 189.9587146937847, "requests": 200, "warmup_excluded": 3 }, "questions_per_request": { "mean": 1.0, "p50": 1.0, "max": 1 }, "ms_per_question": { "p50": 174.91745902225375, "p95": 189.9587146937847, "overall": 172.57170106749982 }, "latency_ms": { "p50": 174.91745902225375, "p95": 189.9587146937847, "requests": 200, "warmup_excluded": 3 }, "control": "none", "model": "devision-0.1", "temperature": { "choice": 1.5922062397003174, "noul": 1.825764536857605, "by_options": { "choice:11+": 1.201171636581421, "choice:2": 1.9487814903259277, "choice:3-5": 1.7575252056121826, "noul:2": 1.825764536857605 } }, "grouping": { "max_questions_per_request": 1, "by": "none (one question per request)" }, "environment": { "threads": 4, "machine": "arm64", "system": "Darwin" }, "run": { "checkpoint": "runs/r15-vision/stage2", "data": "data/r12/cpu_latency.jsonl", "data_sha256": "352f0b4676aeb7389b45a93086169096d6b8e3962568543cfa373877c3e84020", "role": "unspecified", "device": "cpu", "temperature": "fitted", "control": "none", "seed": 0, "code": "6feb0b3+dirty" } }