Datasets:

RobotsMali
/

finBamSpeech

text stringlengths 8 118	speakerID int64 9k 9k	split stringclasses 1 value
milyɔn kelen juru sara waati ka teli yan.	9,004	train
Karidantɛ be filɛ kɔnti dayɛlɛli kama.	9,004	train
Musaka ka ca kalo in na.	9,004	train
Juru misɛnnin musaka fosi tɛ dɔ fara i ka jago liɲi kan.	9,004	train
Apilikasɔn tɛ sɔn ka milyɔn saba ci ni gundo taamasiyɛn tɛ Mɔpti.	9,004	train
Bɔlɔlɔsira jago fɛ, milyɔn tan bilara jago kɛlaw ka bolo kan bi.	9,004	train
Bɔlɔlɔsira jago ba tɔ i bɛ wari sɔrɔ.	9,004	train
Jatɛ fɔlɔ lakanana koɲuman.	9,004	train
Apilikasɔn kura in lakanani sariya b'a jira ko gundo taamasiyɛn dɔ.	9,004	train
Juru bɔ.	9,004	train
Sugu mɔgɔw bɛ be apilikasɔn matrafa.	9,004	train
Kariti kɔnɔna.	9,004	train
waa tan ni duuru bɔra azansi la teliya la sugu liɲi kan bi.	9,004	train
Kariti nimɔrɔ donni bɛ ɲini bɔlɔlɔsira fɛ walasa ka wari sɔrɔ.	9,004	train
Kɔnti sɛgɛsɛgɛli bɛ kɛ kalo o kalo kunfɛsugandili sariya fɛ.	9,004	train
Ni i ɲinɛna i ka gundo taamasiyɛn kɔ, taga azansi la sisan.	9,004	train
Wari cili musaka bɛ tigɛ i ka milyɔn bi saba la.	9,004	train
Ne bɛ se ka n ka wari lakana bɔlɔlɔsira kɔnti fɛ bi wa?	9,004	train
Telefɔni nimɔrɔ bɛɛ bɛ sɔrɔ bɔlɔlɔsira kɔnti kan su ni tile.	9,004	train
Ne ka kan ka kɔnti sariya labato i n'a fɔ Taamasɛbɛn jira.	9,004	train
Kɔnti dayɛlɛni ko la, dɛmɛ ɲini bɔlɔlɔsira kan.	9,004	train
Kɔnti dayɛlɛni fɛɛrɛ lakurayali b'a jira ko wari bɛ sugu fereli la.	9,004	train
N bɛ se ka jurumisɛnnin sara sisan sugu jago kama.	9,004	train
Telefɔni nimɔrɔ yɛlɛmani bɛ kɛ fu azansi la ni karidantɛ ye.	9,004	train
Telefɔni nimɔrɔ kɔrɔ tɛ baara kɛ kɔnti kura kan.	9,004	train
Kɔnti labaloli bɛ kɛ kalo o kalo kunfɛsugandili sariya fɛ.	9,004	train
Kɔnti dayɛlɛni fɛɛrɛ lakurayali b'a jira ko jago bɛ ka yiriwa sisan.	9,004	train
Kɔnti kura in baara kɛcogo kaɲi kosɛbɛ sugu mɔgɔw fɛ bi.	9,004	train
Apilikasɔn nafamaw sɛbɛnni bɛ kɛ fu ye mɔgɔ bɛɛ fɛ.	9,004	train
milyɔn duuru bɔra telefɔni fɛ ni gundo taamasiyɛn ye Kaye.	9,004	train
Ne bɛ ɲini ka jago kɔnti dayɛlɛ milyɔn tan walasa n ka se ka wari lakanane sɔrɔ.	9,004	train
Kɔnti gundo bɛ lakana sariya fɛ bɔlɔlɔsira kura kan mɔgɔ kuraw bɛɛ kama.	9,004	train
Wari cini musaka misɛnnin ye dɔrɔmɛ tan ye bɔlɔlɔsira kura kan bi.	9,004	train
Wari jatɛ jira teliya la ka di ne ma walasa a ka se ka dɛmɛ sugu fere la.	9,004	train
waa fila ani kɛmɛ seegin bɛ n ka kariti kura kɔnɔ walasa n ka se ka jago sigi bi.	9,004	train
Ne bɛ sɔn ka waa kelen.fu ni fu ci n bangebaga ma bi Sikasso apilikasɔn fɛ.	9,004	train
Ne bɛ sɔn ka waa saba ani kɛmɛ fila ci n teri ma telefɔni fɛ bi.	9,004	train
Apilikasɔn ba tɔ ne bɛ wari lakanane sɔrɔ waa fila bi.	9,004	train
Wari cili bɛ kɛ bɔlɔlɔsira fɛ ni telefɔni gansan ye.	9,004	train
N bɛ se ka kɔnti dayɛlɛ ni karidantɛ kɔrɔ ye sisan telefɔni fɛ?	9,004	train
Telefɔni nimɔrɔ kɔrɔ tɛ baara kɛ kɔnti kura kan sisan.	9,004	train
Azansi baarakɛla bɛ mɔgɔw wele walasa ka dɛmɛ lase u ma.	9,004	train
Azansi kura dayɛlɛra an ka sugu yɔrɔ la kan dɛmɛ.	9,004	train
Tikɛ kura bɔra azansi la.	9,004	train
Ni ka wari ci ta tɛmɛna milyɔn kan a ci sara kan ka tigɛ ka bɔ a la.	9,004	train
Apilikasɔn bɛ n dɛmɛ ka n ka jago jatɛ ɲɛnabɔ teliya la.	9,004	train
Wari jatɛ jira teliya la ka di ne ma sugu kɔnɔn.	9,004	train
Kɔnti dayɛlɛni ko la, a ɲini ka dɛmɛ sɔrɔ bɔlɔlɔsira kan.	9,004	train
Kɔnti gansan.	9,004	train
Kɔnti kura in baara kɛcogo ka ɲi kosɛbɛ.	9,004	train
Ne bɛ na milyɔn kelen ci n bangebaga ma Kaye.	9,004	train
Kɔnti labaloli musaka misɛnnin ye dɔrɔmɛ tan ni duuru ye bɔlɔlɔsira kan.	9,004	train
milyɔn saba bɔra n ka kariti kan sugu baara fɛ bi Bamako.	9,004	train
Waa kelen ani kɛmɛ wɔɔrɔ bɔra n ka kɔnti la sani fɛ.	9,004	train
Telefɔni nimɔrɔ kɔrɔ tɛ baara kɛ kɔnti kura kan.	9,004	train
Wari jatɛ bɛ lajɛ fu telefɔni kan sugu mɔgɔw dɛmɛli kama .	9,004	train
Juru daɲɛ.	9,004	train
Jago sariya kura ba fɔ ki ka musaka man ka tɛmɛna milyɔn fila kan azansi la .	9,004	train
Wari cili bɛ kɛ bɔlɔlɔsira fɛ ni telefɔni gansan ye sugu la.	9,004	train
Apilikasɔn gundo taamasiyɛn filɛra ne bolo bi walasa n ka se ka kɔnti sigi ɲɛfɔ ne dɔgoni ye.	9,004	train
Walasa i ka se ki ka Wari jate don i be taga apilikasɔn kan.	9,004	train
Karidantɛ walima Taamasɛbɛn bɛ ɲini ka kɔnti dayɛlɛ la bi telefɔni fɛ sisan.	9,004	train
Kɔnti kan ka lakana sariya fɛ .	9,004	train
Kariti nimɔrɔ donni bɛ ɲini bɔlɔlɔsira fɛ walasa ka juru sara bi.	9,004	train
Kɔnti kura in baara kɛcogo kaɲi kosɛbɛ sugu liɲi kan bi.	9,004	train
Sugu wari.	9,004	train
Kariti kura bɔli musaka ye kɛmɛ duuru ye Kati azansi la bi.	9,004	train
Juru sɔrɔli musaka ye dɔrɔmɛ bi wolonwula ye sisan bɔlɔlɔsira kura kan.	9,004	train
I ka kan i ka gundo taamasiyɛn sigi walasa ka wari lakana.	9,004	train
Sugu mɔgɔw bɛɛ bɛ baara kɛ ni milyɔn saba ye jago kama bi.	9,004	train
N bɛ se ka kunnafoni kuraw sɔrɔ ni telefɔni nimɔrɔ tɛ wa?	9,004	train
Ka wari bɔ i ka kɔnti kɔnɔn nokoya la su ni tile o kadi mɔgɔ bɛɛ ye Mali kɔnɔn bi.	9,004	train
Apilikasɔn tɛ baara kɛ ni wari te kɔnti la.	9,004	train
Juru sɔrɔli bɛ kɛ ni telefɔni nimɔrɔ ye azansi la.	9,004	train
Waa wɔɔrɔ bɔra telefɔni kɔnɔn ni gundo taamasiyɛn numɛro ye.	9,004	train
Kɔnti dayɛlɛsɛbɛn wɛrɛ bɛ ɲini walasa ka i ka se ka milyɔn naani sɔrɔ banki la.	9,004	train
Ne bɛ se ka kɔnti dayɛlɛ ni telefɔni gansan ye azansi kɔnɔn bi jago kama?	9,004	train
Telefɔni nimɔrɔ kunkɔrɔ tɛ sɔn ka baara kɛ kɔnti kura kan sisan azansi la.	9,004	train
Juru sɔrɔli musaka ye kɛmɛ ni bi duuru ye sisan bɔlɔlɔsira kura fɛ azansi la.	9,004	train
Kɔnti labaloli ye dɔrɔmɛ bi wɔɔrɔ ye kalo o kalo an ka banki la.	9,004	train
Ni bi ka Kariti nimɔrɔ don, i bɛ se ka lajɛ ni milyɔn bi saba bɛ kɔnti kɔnɔn.	9,004	train
Kɔnti labaloli musaka ye dɔrɔmɛ tan ni duuru ye bɔlɔlɔsira kan kalo o kalo.	9,004	train
N bɛ sɔn ka waa fila ani kɛmɛ naani ci n teri ma telefɔni fɛ.	9,004	train
Apilikasɔn gundo taamasiyɛn filɛra ne bolo sugu kɔnɔ bi teliya la.	9,004	train
Apilikasɔn nafamaw bɛ mɔgɔ bɛɛ dɛmɛ ka garisɛgɛ kura lakodɔ.	9,004	train
Bɔlɔlɔsira jago fɛ, milyɔn duuru bɛ mɔgɔw cama bɔlɔ bi Sɛgu.	9,004	train
Kariti nimɔrɔ donni bɛ ɲini bɔlɔlɔsira fɛ walasa ka bɛ lafia.	9,004	train
Apilikasɔn taamasiyɛn bɛ filɛ n bolo waati bɛɛ.	9,004	train
Apilikasɔn bɛ mɔgɔ bɛɛ dɛmɛ ka milyɔn bi saba jatɛbɔ teliya la Mɔpti.	9,004	train
Kɔnti dayɛlɛsɛbɛn bɛ ɲɛ ni i bolo digira tɔgɔsɛbɛn kan.	9,004	train
ka wari jatɛ jira teliya la bɛ n dɛmɛ ka jago sigi.	9,004	train
waa fila ani kɛmɛ fila bɔra n ka kɔnti la sugu baara kama bi Kati la.	9,004	train
sariya labato.	9,004	train
Azansi baarakɛla bɛ wari jatɛ jira ne la kalo o kalo.	9,004	train
waa saba bɛ nɛ ka kɔnti kɔnɔn.	9,004	train
N bɛ se ka kɔnti dayɛlɛ ni telefɔni gansan ye bi wa?	9,004	train
Ne bɛ se ka n ka wari jatɛ lakodɔn sisan teliyala azansi la.	9,004	train
Apilikasɔn ye n dɛmɛ kosɛbɛ n sera ka milyɔn duuru sɔrɔ n ka jago fɛ bi.	9,004	train
Apilikasɔn tɛ baara kɛ ni intɛrinɛti tɛ.	9,004	train
Waa fila ani kɛmɛ seegin bɛ n ka kariti kura kɔnɔ ka jago sigi.	9,004	train

End of preview. Expand in Data Studio

FinBamSpeech

FinBamSpeech is a small, higher-quality, single-speaker corpus of 800 read Bambara sentences about finance, banking, and financial technology. RobotsMali created it to test domain adaptation of its first Bambara VITS checkpoints.

The dataset's narrow domain and single voice make it useful for controlled exploratory fine-tuning, but 800 utterances are not enough to claim broad language, speaker, or topic coverage.

Quick facts

Item	Value
Train examples	750
Test examples	50
Speakers	1
Speech style	Read sentences
Domain	Finance, banking, and FinTech
Audio	Embedded WAV audio
Approximate packaged size	324 MB

Load the dataset

from datasets import load_dataset

dataset = load_dataset("RobotsMali/finBamSpeech")
sample = dataset["train"][0]
print(sample["text"], sample["speakerID"])
decoded = sample["audio"].get_all_samples()
print(decoded.sample_rate, decoded.data.shape)

Fields

text (string): the Bambara sentence read by the speaker
audio (audio): decoded audio and its sampling rate/path information
speakerID (int64): constant identifier for the single speaker; it is not a public identity
split (string): source split label, duplicating the Dataset split organization

Intended uses

Experimental single-speaker Bambara TTS fine-tuning
Research on adaptation to finance-related vocabulary and read speech
Reproduction and comparison of RobotsMali's bam-vits-fintech checkpoints
Qualitative study of plain Bambara versus pseudo-IPA input

This dataset does not certify a model for financial advice, banking transactions, customer support, accessibility, or other high-stakes use.

Limitations and responsible use

Only one speaker is represented; voice, accent, age, and demographic diversity cannot be inferred.
The domain is deliberately narrow and may not generalize to conversational or general-purpose Bambara.
Read sentences do not capture spontaneous speech, dialogue, code-switching, or natural interaction.
The train/test split is very small, and repeated sentence patterns may make results unstable.
Higher quality than afvoices-notag is a relative description, not a published studio-quality measurement.
No published audit covers pronunciation, transcription accuracy, acoustic conditions, speaker privacy, or representativeness.

Do not attempt to identify or impersonate the speaker. Evaluate memorization and voice similarity in any derived model, disclose synthetic speech, and obtain appropriate review before deployment.

Models trained on this dataset

Each model received only 150 additional epochs with batch size 80 on top of an already undertrained base. They remain research checkpoints and have no formal perceptual evaluation.

Attribution

When using the dataset, cite it as RobotsMali/finBamSpeech, link to this repository, state the version/commit used, and describe any filtering or normalization. Questions are welcome in the dataset's Community tab.

Downloads last month: 89

Models trained or fine-tuned on RobotsMali/finBamSpeech

RobotsMali/bam-vits-fintech

Text-to-Audio • 39.6M • Updated about 1 month ago • 21

RobotsMali/bam-vits-pseudo-ipa-fintech

Text-to-Audio • 39.6M • Updated about 1 month ago • 14

Collection including RobotsMali/finBamSpeech

BamVits

Collection

8 items • Updated Sep 3