text
stringlengths
0
15.2k
baseline: 98.8% average pass rate over 1403 tests
math : 85.4% average pass rate over 3385 tests
order : 75.7% average pass rate over 1061 tests
present : 66.0% average pass rate over 721 tests
table : 89.2% average pass rate over 1020 tests
Results by JSONL file:
arxiv_math.jsonl : 85.7% (2507/2927 tests)
baseline : 98.8% (1377/1394 tests)
headers_footers.jsonl : 90.7% (689/760 tests)
long_tiny_text.jsonl : 80.5% (356/442 tests)
multi_column.jsonl : 83.8% (741/884 tests)
old_scans.jsonl : 46.2% (243/526 tests)
old_scans_math.jsonl : 83.6% (383/458 tests)
table_tests.jsonl : 89.2% (912/1022 tests)
dots.mocr-mxfp8 : Average Score: 82.5% ± 0.9% (average of per-JSONL scores)
absent : 90.2% average pass rate over 823 tests
baseline: 99.0% average pass rate over 1403 tests
math : 86.6% average pass rate over 3385 tests
order : 76.4% average pass rate over 1061 tests
present : 65.6% average pass rate over 721 tests
table : 88.2% average pass rate over 1020 tests
Results by JSONL file:
arxiv_math.jsonl : 87.0% (2546/2927 tests)
baseline : 99.0% (1380/1394 tests)
headers_footers.jsonl : 90.1% (685/760 tests)
long_tiny_text.jsonl : 79.9% (353/442 tests)
multi_column.jsonl : 84.7% (749/884 tests)
old_scans.jsonl : 46.8% (246/526 tests)
old_scans_math.jsonl : 84.1% (385/458 tests)
table_tests.jsonl : 88.3% (902/1022 tests)
dots.mocr-nvfp4 : Average Score: 82.2% ± 0.9% (average of per-JSONL scores)
absent : 91.7% average pass rate over 823 tests
baseline: 98.8% average pass rate over 1403 tests
math : 85.3% average pass rate over 3385 tests
order : 76.4% average pass rate over 1061 tests
present : 63.8% average pass rate over 721 tests
table : 89.3% average pass rate over 1020 tests
Results by JSONL file:
arxiv_math.jsonl : 85.7% (2507/2927 tests)
baseline : 98.8% (1377/1394 tests)
headers_footers.jsonl : 91.8% (698/760 tests)
long_tiny_text.jsonl : 77.6% (343/442 tests)
multi_column.jsonl : 84.7% (749/884 tests)
old_scans.jsonl : 46.2% (243/526 tests)
old_scans_math.jsonl : 83.4% (382/458 tests)
table_tests.jsonl : 89.3% (913/1022 tests)
Cleaning up <playwright._impl._playwright.Playwright object at 0x106d477a0>
Cleaning up <playwright._impl._playwright.Playwright object at 0x108745310>
Cleaning up <playwright._impl._playwright.Playwright object at 0x106f4b860>
Cleaning up <playwright._impl._playwright.Playwright object at 0x106d4cdd0>
Cleaning up <playwright._impl._playwright.Playwright object at 0x1086f3050>
Cleaning up <playwright._impl._playwright.Playwright object at 0x106f49790>
Cleaning up <playwright._impl._playwright.Playwright object at 0x1086cb0b0>
Cleaning up <playwright._impl._playwright.Playwright object at 0x10861c410>