Datasets:
license: mit
task_categories:
- text-to-speech
dataset_info:
features:
- name: audio
dtype: audio
- name: transcript
dtype: string
- name: language_code
dtype: string
splits:
- name: train
num_bytes: 508120568184.992
num_examples: 736272
download_size: 597640766127
dataset_size: 508120568184.992
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
The Dataset associated with the Paper "Meta Learning Text-to-Speech Synthesis in over 7000 Languages" by Florian Lux, Sarina Meyer, Lyonel Behringer, Frank Zalkow, Phat Do, Matt Coler, Emanuël A. P. Habets and Ngoc Thang Vu (Interspeech 2024).
We generate 2000 spoken utterances per language using the subsets of the eBible dataset [1] that are under free licenses as the text input to the MMS TTS models [2].
The languages associated with the following ISO-639-3 codes are represented in this dataset:
acf, bss, deu, inb, nca, quh, wap, acr, bus, dgr, ind, maz, nch, qul, tav, wmw, acu, byr, dik, iou, mbb, ncj, qvc, tbc, xed, agd, bzh, djk, ipi, mbc, ncl, qve, tbg, xon, agg, bzj, dop, jac, mbh, ncu, qvh, tbl, xtd, agn, caa, jic, mbj, ndj, qvm, tbz, xtm, agr, cab, emp, jiv, mbt, nfa, qvn, tca, yaa, agu, cap, eng, jvn, mca, ngp, qvs, tcs, yad, aia, car, ese, mcb, ngu, qvw, yal, cax, kaq, mcd, nhe, qvz, tee, ycn, ake, cbc, far, mco, qwh, yka, alp, cbi, fra, kdc, mcp, nhu, qxh, ame, cbr, gai, kde, mcq, nhw, qxn, tew, yre, amf, cbs, gam, kdl, mdy, nhy, qxo, tfr, yva, amk, cbt, geb, kek, med, nin, rai, zaa, apb, cbu, glk, ken, mee, nko, rgu, zab, apr, cbv, meq, nld, tgo, zac, arl, cco, gng, kje, met, nlg, rop, tgp, zad, grc, klv, mgh, nnq, rro, zai, ata, cek, gub, kmu, mib, noa, ruf, tna, zam, atb, cgc, guh, kne, mie, not, rug, tnk, zao, atg, chf, knf, mih, npl, rus, tnn, zar, awb, chz, gum, knj, mil, sab, tnp, zas, cjo, guo, ksr, mio, obo, seh, toc, zav, azg, cle, gux, kue, mit, omw, sey, tos, zaw, azz, cme, gvc, kvn, miz, ood, sgb, tpi, zca, bao, cni, gwi, kwd, mkl, shp, tpt, zga, bba, cnl, gym, kwf, mkn, ote, sja, trc, ziw, bbb, cnt, gyr, kwi, mop, otq, snn, ttc, zlm, cof, hat, kyc, mox, pab, snp, tte, zos, bgt, con, kyf, mpm, pad, som, tue, zpc, bjr, cot, heb, kyg, mpp, soy, tuf, zpl, bjv, cpa, kyq, mpx, pao, spa, tuo, zpm, bjz, cpb, hlt, kyz, mqb, pib, spp, tur, zpo, bkd, cpu, hns, lac, mqj, pir, spy, txq, zpu, blz, crn, hto, lat, msy, pjt, sri, txu, zpz, bmr, cso, hub, lex, mto, pls, srm, udu, ztq, bmu, ctu, lgl, muy, poi, srn, ukr, zty, bnp, cuc, lid, mxb, pol, stp, upv, zyp, boa, cui, huu, mxq, por, sus, ura, boj, cuk, huv, llg, mxt, poy, suz, urb, box, cwe, hvn, prf, swe, urt, bpr, cya, ign, lww, myk, ptu, swh, usp, bps, daa, ikk, maj, myy, sxb, vid, bqc, dah, nab, qub, tac, vie, bqp, ded, imo, maq, nas, quf, taj, vmy
[1] V. Akerman, D. Baines, D. Daspit, U. Hermjakob et al., “The eBible Corpus: Data and Model Benchmarks for Bible Translation for Low-Resource Languages,” arXiv:2304.09919, 2023.
[2] V. Pratap, A. Tjandra, B. Shi, P. Tomasello, A. Babu, S. Kundu, A. Elkahky, Z. Ni et al., “Scaling speech technology to 1,000+ languages,” Journal of Machine Learning Research, 2024.