fb35f80006640f696a2b8bf8a4ae90eda0c38a32
base v7

 

challenge
Punctuation prediction from conversational language
submitter
Oskar Bujacz
submitted
2023-03-13 21:32:46.397776 UTC
file basename
out

test-B / 18abe2b0ce5d3500cda3ce9147090f94ff9e709f
Metric Score
Weighted-F1 82.75
Hyphens-F1 100.00
Comma-F1 83.15
Ellipsis-F1 56.06
Fullstop-F1 91.64
QMark-F1 81.84
Colon-F1 100.00
Excl-F1 16.67

test-A / 69baa3ceb1cd66034787f27562f1599948bef3d6
Metric Score
Weighted-F1 78.75
Hyphens-F1 100.00
Comma-F1 81.31
Ellipsis-F1 47.23
Fullstop-F1 91.01
QMark-F1 86.84
Colon-F1 100.00
Excl-F1 18.18

dev-0 / 6dc9a567993517b9edb7e673ead4c712e0a55faf
Metric Score
Weighted-F1 85.60
Hyphens-F1 100.00
Comma-F1 85.25
Ellipsis-F1 44.96
Fullstop-F1 95.39
QMark-F1 89.44
Colon-F1 100.00
Excl-F1 0.00

worst items

note: the gold standard is taken from the challenge
# input expected output actual output test-B Weighted-F1
1 audio/cbiz_tc_557.wav 10599a11-3bff-4ab0-836d-f9ed7a7cd8aa Wie:119580-119670 pan:119700-119760 co:119790-119880 O O B-ELLIPSIS O O B-QMARK 0.00
2 audio/cbiz_tc_557.wav 10599a11-3bff-4ab0-836d-f9ed7a7cd8aa @interia.pl:302610-302610 B-FULLSTOP B-FULLSTOPO 0.00
3 audio/cbiz_tc_557.wav af63e562-a918-42cd-a6b7-3f7f7c0111fb Tak:52320-52410 więc:52440-52560 poproszę:52770-53460 O O B-ELLIPSIS B-COMMA O B-FULLSTOP 0.00
4 audio/cbiz_tc_557.wav af63e562-a918-42cd-a6b7-3f7f7c0111fb Mhm:128910-128970 B-QMARK B-FULLSTOP 0.00
5 audio/cbiz_tc_557.wav af63e562-a918-42cd-a6b7-3f7f7c0111fb Czy:196950-196980 ma:197040-197070 pan:197130-197220 jakieś:197280-197490 O O O B-ELLIPSIS O O O O 0.00
6 audio/cbiz_tc_557.wav af63e562-a918-42cd-a6b7-3f7f7c0111fb Mhm:202080-202110 B-QMARK B-FULLSTOP 0.00
7 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 No:19560-19590 B-XCL B-FULLSTOP 0.00
8 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 A:52020-52020 właśnie:52110-52290 to:52320-52410 mój:52470-52710 O O O B-FULLSTOP O B-COMMA O B-ELLIPSIS 0.00
9 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 To:107670-107700 co:107880-107940 to:108000-108030 są:108120-108360 Z:108510-10851… O O O B-ELLIPSIS O O O O B-ELLIPSIS O B-COMMA O B-QMARK O O O O B-COMMA 0.00
10 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 No:144030-144060 ja:145440-145470 O B-ELLIPSIS O B-FULLSTOP 0.00
11 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 O:164400-164400 Boże:164520-164820 O B-FULLSTOP O B-XCL 0.00
12 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 pensję:173730-174120 B-FULLSTOP B-QMARK 0.00
13 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 Czyli:174750-174870 dostanę:174900-175290 ją:175410-175590 Zaraz:175680-175860 pan… O O B-ELLIPSIS O O B-FULLSTOP O O B-QMARK O O B-COMMA 0.00
14 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 O:216540-216540 Boże:216660-216900 Czyli:217350-217470 O B-FULLSTOP B-ELLIPSIS O B-XCL O 0.00
15 audio/cbiz_tc_26.wav a6c94040-593e-4280-8cbf-fb8fb1360e44 Ta:72960-73020 B-ELLIPSIS B-FULLSTOP 0.00
16 audio/cbiz_tc_26.wav a6c94040-593e-4280-8cbf-fb8fb1360e44 to:115170-115230 B-ELLIPSIS O 0.00
17 audio/cbiz_tc_26.wav a6c94040-593e-4280-8cbf-fb8fb1360e44 Termin:122520-122820 płatności:122970-123660 minął:123810-124290 O O B-ELLIPSIS O O B-FULLSTOP 0.00
18 audio/cbiz_tc_26.wav a6c94040-593e-4280-8cbf-fb8fb1360e44 Wartość:155340-155640 samej:155670-155940 O B-ELLIPSIS O O 0.00
19 audio/cbiz_tc_26.wav a6c94040-593e-4280-8cbf-fb8fb1360e44 No:203430-203490 B-ELLIPSIS B-FULLSTOP 0.00
20 audio/cbiz_tc_41.wav 1f0220c3-4246-4648-bbe4-53fcc8fe5318 ta:273000-273060 która:273120-273330 jest:273390-273510 uregulowania:273630-274410… O O O B-FULLSTOP B-COMMA O O B-ELLIPSIS 0.00
21 audio/cbiz_tc_24.wav c431c029-b346-4475-a95a-41139fdfac47 W:135690-135690 tym:135780-135840 momencie:135930-136290 O O B-ELLIPSIS O O B-COMMA 0.00
22 audio/cbiz_tc_24.wav c431c029-b346-4475-a95a-41139fdfac47 od:141270-141330 tej:141450-141510 decyzji:141570-141900 na:141990-142020 adres:14… O O O O O B-ELLIPSIS O O O O O B-FULLSTOP 0.00
23 audio/cbiz_tc_56.wav 1b68cc28-468b-45e7-8d07-41101491b69e tak:151530-151680 B-FULLSTOP B-QMARK 0.00
24 audio/cbiz_tc_56.wav 1b68cc28-468b-45e7-8d07-41101491b69e Ha:154800-154800 B-QMARK B-FULLSTOP 0.00
25 audio/cbiz_tc_56.wav 1b68cc28-468b-45e7-8d07-41101491b69e Tak:154830-154950 B-ELLIPSIS B-FULLSTOP 0.00
26 audio/cbiz_tc_56.wav b1ac6642-6e1d-4b46-83b2-710dc8ad743e Czyli:90480-90600 muszę:100500-100800 jakby:100890-101190 zalogować:101250-101550 … O O O O O B-QMARK O O B-COMMA O O B-FULLSTOP 0.00
27 audio/cbiz_tc_56.wav b1ac6642-6e1d-4b46-83b2-710dc8ad743e Czy:132540-132600 z:132660-132660 O B-ELLIPSIS O O 0.00
28 audio/cbiz_tc_56.wav b1ac6642-6e1d-4b46-83b2-710dc8ad743e Mm:179940-179940 B-ELLIPSIS B-FULLSTOP 0.00
29 audio/cbiz_tc_64.wav 9e40dee7-153b-421b-ae6e-f5504115b826 Hasło:77430-77850 B-QMARK B-FULLSTOP 0.00
30 audio/cbiz_tc_64.wav 9e40dee7-153b-421b-ae6e-f5504115b826 To:123480-123660 znaczy:123930-124260 O B-ELLIPSIS O B-QMARK 0.00
31 audio/cbiz_tc_58.wav 7e32d310-c0b7-4a3d-82b0-9c58cf728b4e problemów:25620-25950 finansowych:26010-26580 O B-FULLSTOP O B-QMARK 0.00
32 audio/cbiz_tc_58.wav 7e32d310-c0b7-4a3d-82b0-9c58cf728b4e Numer:31680-31890 klienta:32040-32610 to:32790-32850 jest:33060-33300 O B-COMMA O B-ELLIPSIS O O O B-QMARK 0.00
33 audio/cbiz_tc_64.wav ed4ff12e-a462-4c82-a004-38ab9fe618bf Zależy:168390-168540 na:168570-168600 jaką:168630-168720 infolinię:168750-168990 a… O O O B-COMMA O O O O O O O B-ELLIPSIS B-FULLSTOP O O O B-COMMA O O O O O O O B-COMMA B-ELLIPSIS 22.22
34 audio/cbiz_tc_64.wav ed4ff12e-a462-4c82-a004-38ab9fe618bf I:49470-49470 też:49500-49560 od:49590-49620 razu:49680-49950 jak:49980-50040 będę… O O O O O O O O O B-COMMA O O O O B-ELLIPSIS O O O B-COMMA O O O O O B-COMMA O B-COMMA O O O 25.00
35 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 Cztery:94770-95010 Ale:95100-95190 dlaczego:95250-95520 aż:95580-95610 tak:95700-9… B-ELLIPSIS O O O O B-QMARK O O O O O B-COMMA O O O B-FULLSTOP O O O O O B-QMARK B-FULLSTOP O O O O B-COMMA O O O O O B-COMMA B-COMMA O O B-QMARK O O O O O B-QMARK 30.00
36 audio/cbiz_tc_557.wav 10599a11-3bff-4ab0-836d-f9ed7a7cd8aa Dwa:64110-64260 cztery:64290-64620 B-COMMA B-FULLSTOP B-COMMA B-COMMA 33.33
37 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 Numer:33990-34200 klienta:34320-34710 to:34800-34890 dwa:34980-35160 dwa:35220-354… O O O B-COMMA B-ELLIPSIS O O O B-COMMA B-COMMA 33.33
38 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 O:91380-91380 matko:91470-91740 To:91800-91830 ja:91890-91920 nie:91980-92010 mam:… O B-XCL O O O O O O O B-FULLSTOP O B-FULLSTOP O O O O O O O B-FULLSTOP 33.33
39 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 No:166200-166230 dobrze:166290-166440 to:166530-166590 co:166680-166740 ja:166770-… O B-COMMA O O O O O B-ELLIPSIS O B-COMMA O O O O O B-COMMA 33.33
40 audio/cbiz_tc_26.wav 597f8543-06b5-4d46-b85c-cd8f43df2264 w:219930-219930 tej:219990-220080 ciemności:220110-220620 dopóki:220680-220950 nie… O O B-COMMA O O B-QMARK O O B-XCL O O B-COMMA O O B-FULLSTOP O O B-FULLSTOP 33.33

Compare with other submission