tiny-hinglish-turn-detector / reports /partial_tinytcn_failures.json
suvradeepp's picture
Publish Tiny Hinglish Turn Detector development preview
35d483e verified
Raw
History Blame Contribute Delete
13.2 kB
{
"counts": {
"examples": 326,
"false_interruptions": 3,
"missed_ends": 150
},
"development_only": true,
"failure_counts_by_slice": {
"dataset": {
"chirp3_1": {
"count": 177,
"false_interruptions": 2,
"missed_ends": 78
},
"chirp3_2": {
"count": 85,
"false_interruptions": 1,
"missed_ends": 35
},
"human_5": {
"count": 5,
"false_interruptions": 0,
"missed_ends": 2
},
"liva_1": {
"count": 36,
"false_interruptions": 0,
"missed_ends": 23
},
"midcentury_1": {
"count": 10,
"false_interruptions": 0,
"missed_ends": 5
}
},
"duration_bin": {
"2-4s": {
"count": 39,
"false_interruptions": 1,
"missed_ends": 15
},
"4-8s": {
"count": 148,
"false_interruptions": 1,
"missed_ends": 68
},
">=8s": {
"count": 136,
"false_interruptions": 1,
"missed_ends": 65
}
},
"filler_type": {
"endfiller": {
"count": 64,
"false_interruptions": 0,
"missed_ends": 0
},
"midfiller": {
"count": 134,
"false_interruptions": 2,
"missed_ends": 58
},
"no_filler": {
"count": 65,
"false_interruptions": 1,
"missed_ends": 55
},
"unknown": {
"count": 63,
"false_interruptions": 0,
"missed_ends": 37
}
},
"language": {
"ara": {
"count": 8,
"false_interruptions": 0,
"missed_ends": 2
},
"ben": {
"count": 9,
"false_interruptions": 1,
"missed_ends": 3
},
"dan": {
"count": 9,
"false_interruptions": 0,
"missed_ends": 7
},
"deu": {
"count": 13,
"false_interruptions": 0,
"missed_ends": 6
},
"eng": {
"count": 76,
"false_interruptions": 0,
"missed_ends": 43
},
"fin": {
"count": 11,
"false_interruptions": 0,
"missed_ends": 4
},
"fra": {
"count": 16,
"false_interruptions": 0,
"missed_ends": 7
},
"hin": {
"count": 16,
"false_interruptions": 0,
"missed_ends": 7
},
"ind": {
"count": 11,
"false_interruptions": 0,
"missed_ends": 3
},
"ita": {
"count": 7,
"false_interruptions": 0,
"missed_ends": 4
},
"jpn": {
"count": 9,
"false_interruptions": 0,
"missed_ends": 3
},
"kor": {
"count": 10,
"false_interruptions": 0,
"missed_ends": 7
},
"mar": {
"count": 8,
"false_interruptions": 0,
"missed_ends": 3
},
"nld": {
"count": 15,
"false_interruptions": 0,
"missed_ends": 7
},
"nor": {
"count": 10,
"false_interruptions": 0,
"missed_ends": 4
},
"pol": {
"count": 10,
"false_interruptions": 0,
"missed_ends": 3
},
"por": {
"count": 14,
"false_interruptions": 0,
"missed_ends": 6
},
"rus": {
"count": 14,
"false_interruptions": 0,
"missed_ends": 5
},
"spa": {
"count": 21,
"false_interruptions": 1,
"missed_ends": 8
},
"tur": {
"count": 9,
"false_interruptions": 1,
"missed_ends": 5
},
"ukr": {
"count": 7,
"false_interruptions": 0,
"missed_ends": 5
},
"vie": {
"count": 12,
"false_interruptions": 0,
"missed_ends": 4
},
"zho": {
"count": 11,
"false_interruptions": 0,
"missed_ends": 4
}
},
"synthetic": {
"False": {
"count": 55,
"false_interruptions": 0,
"missed_ends": 33
},
"True": {
"count": 271,
"false_interruptions": 3,
"missed_ends": 117
}
}
},
"highest_confidence_false_interruptions": [
{
"case_id": "case_5c015a40de99",
"dataset": "chirp3_2",
"duration_bin": "2-4s",
"filler_type": "no_filler",
"language": "ben",
"p_end": 0.7609416246414185,
"predicted": "END",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "HOLD"
},
{
"case_id": "case_3b6f208b1e8f",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "midfiller",
"language": "tur",
"p_end": 0.7598459124565125,
"predicted": "END",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "HOLD"
},
{
"case_id": "case_9d30fab8fc0c",
"dataset": "chirp3_1",
"duration_bin": ">=8s",
"filler_type": "midfiller",
"language": "spa",
"p_end": 0.7539693117141724,
"predicted": "END",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "HOLD"
}
],
"highest_confidence_missed_ends": [
{
"case_id": "case_b489d01be9a9",
"dataset": "chirp3_2",
"duration_bin": "4-8s",
"filler_type": "midfiller",
"language": "mar",
"p_end": 0.09413643926382065,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_422104ea5f81",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "no_filler",
"language": "zho",
"p_end": 0.1314292997121811,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_132fd41ebe62",
"dataset": "liva_1",
"duration_bin": ">=8s",
"filler_type": "unknown",
"language": "eng",
"p_end": 0.14234659075737,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": false,
"target": "END"
},
{
"case_id": "case_64757b5114d0",
"dataset": "chirp3_1",
"duration_bin": ">=8s",
"filler_type": "midfiller",
"language": "zho",
"p_end": 0.14669735729694366,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_8ae5f513d226",
"dataset": "chirp3_2",
"duration_bin": "4-8s",
"filler_type": "midfiller",
"language": "dan",
"p_end": 0.16363245248794556,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_a25f5a8186ea",
"dataset": "chirp3_2",
"duration_bin": ">=8s",
"filler_type": "no_filler",
"language": "nor",
"p_end": 0.17043010890483856,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_7f6311a717ef",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "midfiller",
"language": "ita",
"p_end": 0.19661107659339905,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_a6bd5b2ad407",
"dataset": "midcentury_1",
"duration_bin": "4-8s",
"filler_type": "unknown",
"language": "eng",
"p_end": 0.201650470495224,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": false,
"target": "END"
},
{
"case_id": "case_52e9791a43fd",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "midfiller",
"language": "nld",
"p_end": 0.2067912071943283,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_02dd78a5a5e2",
"dataset": "liva_1",
"duration_bin": ">=8s",
"filler_type": "unknown",
"language": "eng",
"p_end": 0.20779727399349213,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": false,
"target": "END"
},
{
"case_id": "case_5e590f379e3b",
"dataset": "liva_1",
"duration_bin": ">=8s",
"filler_type": "unknown",
"language": "eng",
"p_end": 0.2102150022983551,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": false,
"target": "END"
},
{
"case_id": "case_2f763e1b44ad",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "no_filler",
"language": "kor",
"p_end": 0.21097739040851593,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_c26821c6f9a1",
"dataset": "chirp3_1",
"duration_bin": ">=8s",
"filler_type": "no_filler",
"language": "por",
"p_end": 0.2196618914604187,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_1e0b152035b5",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "no_filler",
"language": "fra",
"p_end": 0.22419698536396027,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_d7dbe32b32d9",
"dataset": "chirp3_1",
"duration_bin": ">=8s",
"filler_type": "no_filler",
"language": "jpn",
"p_end": 0.2265254259109497,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_f3f629c05c08",
"dataset": "chirp3_1",
"duration_bin": "4-8s",
"filler_type": "no_filler",
"language": "deu",
"p_end": 0.23065729439258575,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_7490ba2b899b",
"dataset": "liva_1",
"duration_bin": ">=8s",
"filler_type": "unknown",
"language": "eng",
"p_end": 0.23717747628688812,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": false,
"target": "END"
},
{
"case_id": "case_3ab2138eb24c",
"dataset": "chirp3_2",
"duration_bin": "4-8s",
"filler_type": "no_filler",
"language": "ukr",
"p_end": 0.23719172179698944,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
},
{
"case_id": "case_6dc5afca3bdc",
"dataset": "liva_1",
"duration_bin": ">=8s",
"filler_type": "unknown",
"language": "eng",
"p_end": 0.238896444439888,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": false,
"target": "END"
},
{
"case_id": "case_3b8c7b90dd59",
"dataset": "chirp3_1",
"duration_bin": ">=8s",
"filler_type": "no_filler",
"language": "nld",
"p_end": 0.23906758427619934,
"predicted": "HOLD",
"review_note": "Listen under authorized local access; do not export audio or transcript.",
"synthetic": true,
"target": "END"
}
],
"privacy": "Case IDs are one-way hashes. No audio, transcript, raw record ID, or source path is included. Hypotheses require authorized local listening.",
"scope": "preview-only: 3265 rows from train-00010-of-00083; official test remains sealed",
"split": "validation",
"threshold": 0.741000771522522
}