Buckets:

rhtse's picture
download
raw
1.42 kB
{
"status": "POC: ASR stage runs; alignment stage NOT implemented (parked)",
"clips": 20,
"audio_seconds": 161.64,
"wall_s": 18.2,
"rtf_asr_only": 0.113,
"hourly_label_cost_usd": 0.28,
"next_steps": "espeak reference phonemization of plane-restored text; CTC forced alignment; agreement stats; 2027 campaign gate",
"samples": [
{
"clip": "fleurs-he-000.wav",
"phonemes": "k aɪ z a m ɪ n b oː f ə n i r r a d b r o l x a v ɛ x i p ɛ l ɛ ɡ a m i t b a x a dʒ a v a n i m uː f j a n b eː m ʊ m i ɡ v a n m ə m a t n e t u b a l oː t b eː f a ʃ t uː t v ɛ t ɛ m iː m a b u l t i m ʃ a dʒ a v a n s a n i m ɑ a v eː m ə m b o t n iː m p i l p e l tʃ i l i s u k aː m ʊ m i ɡ v a n t a v l i n eː m ə a m o t i",
"seconds": 16.08
},
{
"clip": "fleurs-he-001.wav",
"phonemes": "n e m t ts a b ɛ x e l i k a ts f o n m y z a r a x i ʃ e l d r o m a f r i k a l a o r x e ɡ v uː l m o z a m b i k b a m i s r a x z i m p a v y b a z a f o n v i ɡ v u l aː d oː m i u n i h a r k eː ɛ m b i h a p a r k ɛ l i u m i k r u ɡ e r k r o k o d a l",
"seconds": 13.32
},
{
"clip": "fleurs-he-002.wav",
"phonemes": "ʃ e ɡ a t e l m i d i n a z a ɡ a b a t x a l a k a n i r a ʃ a n a ʃ i m a j y s o v l a n i v m i v i n i m b eː d i j a ʃ i m i t e l i m b e m i d i n o t a d a ʃ o ts u x i m i t e k l e m",
"seconds": 7.62
}
]
}

Xet Storage Details

Size:
1.42 kB
·
Xet hash:
515e4e2cffe81d5825b2aaac9908be0aa688c2f41fb17cdf69bd1d074c89818f

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.