formospeech/tat_moe_multichannel

Dataset Information

• Corpus: tat_moe_multichannel
• Status: Private and Closed

Subset Stats

Language	condenser	lavalier	ios	android	xyh6_x	xyh6_y
nan-tw	98,091 / 102.33	97,880 / 102.06	94,517 / 98.22	87,338 / 91.80	97,460 / 101.66	96,536 / 100.60
Total	98,091 / 102.33	97,880 / 102.06	94,517 / 98.22	87,338 / 91.80	97,460 / 101.66	96,536 / 100.60

train

Language	condenser	lavalier	ios	android	xyh6_x	xyh6_y
nan-tw	74,746 / 78.61	74,576 / 78.38	71,980 / 75.31	66,972 / 71.05	74,162 / 77.98	73,494 / 77.16
Total	74,746 / 78.61	74,576 / 78.38	71,980 / 75.31	66,972 / 71.05	74,162 / 77.98	73,494 / 77.16

eval

Language	condenser	lavalier	ios	android	xyh6_x	xyh6_y
nan-tw	10,761 / 11.13	10,734 / 11.11	10,109 / 10.50	7,866 / 8.27	10,718 / 11.10	10,459 / 10.85
Total	10,761 / 11.13	10,734 / 11.11	10,109 / 10.50	7,866 / 8.27	10,718 / 11.10	10,459 / 10.85

test

Language	condenser	lavalier	ios	android	xyh6_x	xyh6_y
nan-tw	12,584 / 12.59	12,570 / 12.58	12,428 / 12.41	12,500 / 12.49	12,580 / 12.59	12,583 / 12.59
Total	12,584 / 12.59	12,570 / 12.58	12,428 / 12.41	12,500 / 12.49	12,580 / 12.59	12,583 / 12.59