AfriSpeech/gloss-mt-train-fr: African Glossary-Constrained Translation Dataset (French Source)

AfriSpeech/gloss-mt-train-fr provides full-scale parallel datasets engineered specifically for glossary-constrained machine translation from French into African languages.

Each training example features nominal phrases in the source sentence replaced in-place with their verified target language translation enclosed in <term> markers (e.g. The <kannifoɔ> made a covenant with God...).

• Author: Mich-Seth Owusu (michseth@ghananlp.org)
• Organization: AfriSpeech & Ghana NLP
• Codebase: github.com/AfriSpeech/afrigloss-mt
• Source Language: French (fr)
• Total African Subsets: 402 languages
• Total Parallel Sentences: 4,627,410

---

Supported Language Subsets

Subset (config)	Target Language	Train	Validation	Test	Total Examples
hay	HAY	17,723	984	984	19,691
hbb	HBB	6,251	347	347	6,945
hbn	HBN	5,345	296	296	5,937
hdy	HDY	13,366	742	742	14,850
heh	HEH	5,087	282	282	5,651
her	HER	18,588	1,032	1,032	20,652
hwo	HWO	5,564	309	309	6,182
ibb	IBB	24,649	1,369	1,369	27,387
iby	IBY	5,820	323	323	6,466
ich	ICH	6,197	344	344	6,885
idc	IDC	2,757	153	153	3,063
ife	IFE	6,152	341	341	6,834
ijn	IJN	21,524	1,195	1,195	23,914
ikk	IKK	22,665	1,259	1,259	25,183
ikw	IKW	6,251	347	347	6,945
ikz	IKZ	5,405	300	300	6,005
ilb	ILB	5,498	305	305	6,108
iqw	IQW	22,777	1,265	1,265	25,307
iri	IRI	10,948	608	608	12,164
irk	IRK	27,310	1,517	1,517	30,344
isn	ISN	4,044	224	224	4,492
iyx	IYX	6,107	339	339	6,785
izr	IZR	5,958	331	331	6,620
izz	IZZ	22,455	1,247	1,247	24,949
jab	JAB	6,247	347	347	6,941
jbn	JBN	2,736	151	151	3,038
jgo	JGO	4,664	259	259	5,182
jib	JIB	5,671	314	314	6,299
jid	JID	5,414	300	300	6,014
jit	JIT	8,400	466	466	9,332
jmc	JMC	19,250	1,069	1,069	21,388
jnj	JNJ	1,319	73	73	1,465
jow	JOW	6,488	360	360	7,208
jub	JUB	781	43	43	867
jum	JUM	2,937	163	163	3,263
kaj	KAJ	5,518	306	306	6,130
kbn	KBN	6,062	336	336	6,734
kbo	KBO	7,627	423	423	8,473
kbr	KBR	16,308	905	905	18,118
kby	KBY	3,769	209	209	4,187
kcc	KCC	594	33	33	660
kcg	KCG	6,117	339	339	6,795
kcq	KCQ	5,792	321	321	6,434
kcu	KCU	4,213	233	233	4,679
kdc	KDC	13,330	740	740	14,810
kdh	KDH	6,412	356	356	7,124
kdi	KDI	23,949	1,330	1,330	26,609
kdj	KDJ	23,313	1,295	1,295	25,903
kdl	KDL	5,596	310	310	6,216
ked	KED	4,307	239	239	4,785
ken	KEN	5,454	302	302	6,058
keo	KEO	5,868	326	326	6,520
ker	KER	6,053	336	336	6,725
khy	KHY	5,847	324	324	6,495
kia	KIA	26,238	1,457	1,457	29,152
kik	KIK	57,324	3,184	3,184	63,692
kiv	KIV	2,038	113	113	2,264
kki	KKI	4,145	230	230	4,605
kkj	KKJ	6,141	341	341	6,823
kln	KLN	40,745	2,263	2,263	45,271
klu	KLU	20,451	1,136	1,136	22,723
kma	KMA	6,191	343	343	6,877
kmw	KMW	5,191	288	288	5,767
knf	KNF	8,045	446	446	8,937
knk	KNK	19,268	1,070	1,070	21,408
kno	KNO	6,041	335	335	6,711
kpz	KPZ	16,978	943	943	18,864
kqn	KQN	5,763	320	320	6,403
kqo	KQO	5,855	325	325	6,505
kqp	KQP	6,087	338	338	6,763
kqs	KQS	19,873	1,103	1,103	22,079
kqy	KQY	3,595	199	199	3,993
krs	KRS	7,960	442	442	8,844
krw	KRW	5,664	314	314	6,292
krx	KRX	5,204	289	289	5,782
ksb	KSB	4,954	275	275	5,504
ksf	KSF	4,824	268	268	5,360
ksp	KSP	8,277	459	459	9,195
kss	KSS	16,799	933	933	18,665
ktb	KTB	13,324	740	740	14,804
ktj	KTJ	5,479	304	304	6,087
kub	KUB	8,979	498	498	9,975
kuj	KUJ	21,403	1,188	1,188	23,779
kus	KUS	19,810	1,100	1,100	22,010
kvj	KVJ	21,719	1,206	1,206	24,131
kwm	KWM	4,257	236	236	4,729
kwn	KWN	14,755	819	819	16,393
kxc	KXC	19,481	1,082	1,082	21,645
kya	KYA	5,865	325	325	6,515
kyf	KYF	5,509	306	306	6,121
kyq	KYQ	5,989	332	332	6,653
kzr	KZR	5,487	304	304	6,095
lag	LAG	5,641	313	313	6,267
lai	LAI	19,405	1,077	1,077	21,559
lam	LAM	15,311	850	850	17,011
lap	LAP	5,695	316	316	6,327
ldb	LDB	6,113	339	339	6,791
ldk	LDK	6,012	333	333	6,678
leb	LEB	4,868	270	270	5,408
lee	LEE	5,305	294	294	5,893
lef	LEF	19,008	1,055	1,055	21,118
leh	LEH	19,457	1,080	1,080	21,617
lem	LEM	5,857	325	325	6,507
lgm	LGM	5,234	290	290	5,814
lia	LIA	6,300	350	350	7,000
lik	LIK	6,192	343	343	6,878
lip	LIP	5,855	325	325	6,505
liy	LIY	6,094	338	338	6,770
lla	LLA	4,296	238	238	4,772
llb	LLB	14,190	788	788	15,766
lmd	LMD	5,666	314	314	6,294
lme	LME	7,345	408	408	8,161
lmp	LMP	5,832	323	323	6,478
lnl	LNL	5,970	331	331	6,632
lns	LNS	22,166	1,231	1,231	24,628
lob	LOB	4,975	276	276	5,527
log	LOG	6,118	339	339	6,796
lol	LOL	24,367	1,353	1,353	27,073
lom	LOM	4,747	263	263	5,273
lon	LON	19,391	1,077	1,077	21,545
loq	LOQ	5,744	319	319	6,382
lot	LOT	11,162	620	620	12,402
lsm	LSM	20,286	1,127	1,127	22,540
lth	LTH	6,381	354	354	7,089
lto	LTO	22,062	1,225	1,225	24,512
luc	LUC	14,545	807	807	16,159
lue	LUE	19,826	1,101	1,101	22,028
luj	LUJ	5,559	308	308	6,175
lun	LUN	27,642	1,535	1,535	30,712
lwg	LWG	31,962	1,775	1,775	35,512
lwo	LWO	4,922	273	273	5,468
lyn	LYN	5,516	306	306	6,128
maf	MAF	27,353	1,519	1,519	30,391
mas	MAS	36,519	2,028	2,028	40,575
maw	MAW	5,839	324	324	6,487
mbu	MBU	19,517	1,084	1,084	21,685
mck	MCK	21,005	1,166	1,166	23,337
mcn	MCN	21,949	1,219	1,219	24,387
mcu	MCU	5,418	300	300	6,018
mda	MDA	5,734	318	318	6,370
mdm	MDM	6,009	333	333	6,675
mdy	MDY	17,502	972	972	19,446
men	MEN	5,617	311	311	6,239
meq	MEQ	6,147	341	341	6,829
mev	MEV	5,547	308	308	6,163
mey	MEY	1,090	60	60	1,210
mfg	MFG	7,161	397	397	7,955
mfh	MFH	6,167	342	342	6,851
mfi	MFI	6,624	367	367	7,358
mfk	MFK	6,182	343	343	6,868
mfm	MFM	4,399	244	244	4,887
mfn	MFN	7,457	414	414	8,285
mfo	MFO	6,144	341	341	6,826
mfv	MFV	6,213	345	345	6,903
mfx	MFX	2,717	150	150	3,017
mgc	MGC	2,932	162	162	3,256
mge	MGE	5,220	290	290	5,800
mgg	MGG	5,724	317	317	6,358
mgj	MGJ	2,956	164	164	3,284
mgo	MGO	9,615	534	534	10,683
mgq	MGQ	5,018	278	278	5,574
mgs	MGS	5,417	300	300	6,017
mgv	MGV	11,622	645	645	12,912
mgw	MGW	5,795	321	321	6,437
mgy	MGY	5,736	318	318	6,372
mhd	MHD	1,984	110	110	2,204
mho	MHO	5,790	321	321	6,432
mhw	MHW	4,830	268	268	5,366
mif	MIF	6,946	385	385	7,716
mkl	MKL	20,899	1,161	1,161	23,221
mkw	MKW	22,821	1,267	1,267	25,355
mlw	MLW	6,410	356	356	7,122
mmu	MMU	5,387	299	299	5,985
mmy	MMY	5,847	324	324	6,495
mnf	MNF	4,932	273	273	5,478
mnh	MNH	3,982	221	221	4,424
mnk	MNK	22,716	1,262	1,262	25,240
moa	MOA	6,267	348	348	6,963
moy	MOY	3,684	204	204	4,092
moz	MOZ	6,002	333	333	6,668
mpa	MPA	5,671	314	314	6,299
mpe	MPE	4,909	272	272	5,453
mpg	MPG	21,560	1,197	1,197	23,954
mqb	MQB	6,147	341	341	6,829
msc	MSC	7,489	415	415	8,319
mse	MSE	20,485	1,138	1,138	22,761
mua	MUA	20,733	1,151	1,151	23,035
mug	MUG	21,131	1,173	1,173	23,477
muh	MUH	6,372	354	354	7,080
mur	MUR	6,371	353	353	7,077
muy	MUY	5,970	331	331	6,632
mwe	MWE	5,415	300	300	6,015
mwm	MWM	24,333	1,351	1,351	27,035
mwn	MWN	16,340	907	907	18,154
mxx	MXX	3,606	200	200	4,006
myb	MYB	34,442	1,913	1,913	38,268
mye	MYE	19,188	1,066	1,066	21,320
myk	MYK	5,981	332	332	6,645
myx	MYX	28,966	1,609	1,609	32,184
mzj	MZJ	5,706	316	316	6,338
mzm	MZM	6,202	344	344	6,890
mzw	MZW	20,388	1,132	1,132	22,652
nat	NAT	5,223	290	290	5,803
naw	NAW	6,525	362	362	7,249
nbl	NBL	12,161	675	675	13,511
ncr	NCR	6,022	334	334	6,690
ncu	NCU	5,328	296	296	5,920
nde	NDE	22,789	1,265	1,265	25,319
ndh	NDH	5,381	298	298	5,977
ndi	NDI	5,978	332	332	6,642
ndj	NDJ	5,392	299	299	5,990
ndp	NDP	5,691	316	316	6,323
ndv	NDV	6,690	371	371	7,432
ndz	NDZ	6,093	338	338	6,769
ney	NEY	2,314	128	128	2,570
nga	NGA	16,351	908	908	18,167
ngc	NGC	11,477	637	637	12,751
nge	NGE	6,012	333	333	6,678
ngj	NGJ	6,172	342	342	6,856
ngn	NGN	6,364	353	353	7,070
ngp	NGP	18,364	1,020	1,020	20,404
ngz	NGZ	6,096	338	338	6,772
nhr	NHR	5,943	330	330	6,603
nhu	NHU	5,983	332	332	6,647
nie	NIE	5,780	321	321	6,422
nih	NIH	5,163	286	286	5,735
nim	NIM	4,051	224	224	4,499
nin	NIN	6,041	335	335	6,711
niy	NIY	6,051	336	336	6,723
njd	NJD	3,993	221	221	4,435
nka	NKA	25,021	1,389	1,389	27,799
nkn	NKN	10,034	557	557	11,148
nko	NKO	6,901	383	383	7,667
nla	NLA	5,389	299	299	5,987
nle	NLE	5,923	328	328	6,579
nmq	NMQ	3,544	196	196	3,936
nnb	NNB	21,791	1,210	1,210	24,211
nnq	NNQ	5,386	299	299	5,984
nnw	NNW	5,513	306	306	6,125
nse	NSE	8,201	455	455	9,111
ntk	NTK	5,797	321	321	6,439
ntr	NTR	6,199	344	344	6,887
nuj	NUJ	10,819	601	601	12,021
nus	NUS	18,306	1,017	1,017	20,340
nwb	NWB	5,566	309	309	6,184
nww	NWW	5,522	306	306	6,134
nxd	NXD	4,809	267	267	5,343
nyb	NYB	1,185	65	65	1,315
nyd	NYD	18,669	1,037	1,037	20,743
nyf	NYF	5,375	298	298	5,971
nyj	NYJ	2,372	131	131	2,634
nym	NYM	7,368	409	409	8,186
nyo	NYO	32,815	1,822	1,822	36,459
nyu	NYU	7,821	434	434	8,689
nyy	NYY	29,603	1,644	1,644	32,891
nza	NZA	6,014	334	334	6,682
odu	ODU	5,593	310	310	6,213
okr	OKR	20,906	1,161	1,161	23,228
oku	OKU	5,723	317	317	6,357
old	OLD	18,838	1,046	1,046	20,930
omi	OMI	7,371	409	409	8,189
otr	OTR	5,324	295	295	5,914
oyd	OYD	3,161	175	175	3,511
ozm	OZM	8,477	470	470	9,417
pcm	PCM	24,157	1,341	1,341	26,839
pfe	PFE	5,352	297	297	5,946
pkb	PKB	5,374	298	298	5,970
pko	PKO	20,229	1,123	1,123	22,475
plt	PLT	99,447	5,524	5,524	110,495
pny	PNY	5,890	327	327	6,544
pnz	PNZ	6,080	337	337	6,754
pov	POV	23,491	1,304	1,304	26,099
poy	POY	19,151	1,063	1,063	21,277
puu	PUU	18,524	1,029	1,029	20,582
rag	RAG	34,217	1,900	1,900	38,017
rel	REL	5,541	307	307	6,155
rif	RIF	4,155	230	230	4,615
rim	RIM	5,281	293	293	5,867
rub	RUB	8,073	448	448	8,969
ruc	RUC	4,297	238	238	4,773
ruf	RUF	17,114	950	950	19,014
rwk	RWK	5,030	279	279	5,588
rwm	RWM	1,647	91	91	1,829
saa	SAA	1,154	64	64	1,282
saq	SAQ	6,039	335	335	6,709
sav	SAV	7,184	399	399	7,982
sba	SBA	21,644	1,202	1,202	24,048
sbd	SBD	19,963	1,108	1,108	22,179
sbk	SBK	5,395	299	299	5,993
sbm	SBM	4,565	253	253	5,071
sbp	SBP	5,933	329	329	6,591
sby	SBY	5,194	288	288	5,770
sef	SEF	20,512	1,139	1,139	22,790
ses	SES	18,966	1,053	1,053	21,072
sev	SEV	11,317	628	628	12,573
sfw	SFW	27,091	1,505	1,505	30,101
sgw	SGW	11,580	643	643	12,866
shi	SHI	9,582	532	532	10,646
shk	SHK	20,400	1,133	1,133	22,666
shq	SHQ	1,440	80	80	1,600
shr	SHR	16,439	913	913	18,265
shu	SHU	13,357	741	741	14,839
sig	SIG	5,967	331	331	6,629
sil	SIL	21,738	1,207	1,207	24,152
snf	SNF	4,949	274	274	5,497
snm	SNM	1,692	93	93	1,878
snw	SNW	19,004	1,055	1,055	21,114
som	SOM	21,543	1,196	1,196	23,935
sop	SOP	4,454	247	247	4,948
sor	SOR	3,366	187	187	3,740
soy	SOY	5,109	283	283	5,675
spp	SPP	7,341	407	407	8,155
sqa	SQA	1,627	90	90	1,807
srr	SRR	21,989	1,221	1,221	24,431
ssc	SSC	5,793	321	321	6,435
ssy	SSY	5,229	290	290	5,809
suk	SUK	26,626	1,479	1,479	29,584
suq	SUQ	4,068	226	226	4,520
sus	SUS	20,035	1,112	1,112	22,259
swb	SWB	5,508	305	305	6,118
swk	SWK	21,790	1,210	1,210	24,210
sxb	SXB	5,623	312	312	6,247
szv	SZV	5,895	327	327	6,549
tan	TAN	2,430	134	134	2,698
tap	TAP	3,709	205	205	4,119
taq	TAQ	3,836	213	213	4,262
tbt	TBT	5,779	321	321	6,421
tcc	TCC	25,487	1,415	1,415	28,317
tcd	TCD	5,383	299	299	5,981
ted	TED	5,296	294	294	5,884
tem	TEM	4,896	272	272	5,440
tex	TEX	6,912	384	384	7,680
tga	TGA	5,885	326	326	6,537
tgw	TGW	19,625	1,090	1,090	21,805
thk	THK	5,465	303	303	6,071
thv	THV	2,001	111	111	2,223
tke	TKE	7,201	399	399	7,999
tlj	TLJ	13,713	761	761	15,235
tll	TLL	21,787	1,210	1,210	24,207
tng	TNG	5,905	328	328	6,561
tnr	TNR	5,335	296	296	5,927
tod	TOD	18,027	1,001	1,001	20,029
toh	TOH	24,006	1,333	1,333	26,672
tpm	TPM	5,442	302	302	6,046
tsa	TSA	6,150	341	341	6,832
tsc	TSC	21,645	1,202	1,202	24,049
tsv	TSV	5,382	298	298	5,978
tsw	TSW	20,900	1,161	1,161	23,222
ttq	TTQ	4,370	242	242	4,854
ttr	TTR	5,877	326	326	6,529
tui	TUI	41,904	2,327	2,327	46,558
tul	TUL	5,254	291	291	5,836
tuv	TUV	22,186	1,232	1,232	24,650
tvs	TVS	6,944	385	385	7,714
tvu	TVU	5,978	332	332	6,642
tyx	TYX	5,599	311	311	6,221
udu	UDU	9,408	522	522	10,452
urh	URH	23,500	1,305	1,305	26,110
uth	UTH	4,248	236	236	4,720
utr	UTR	6,195	344	344	6,883
vag	VAG	20,683	1,149	1,149	22,981
vai	VAI	4,114	228	228	4,570
ver	VER	3,389	188	188	3,765
vid	VID	16,546	919	919	18,384
vif	VIF	5,809	322	322	6,453
vun	VUN	5,139	285	285	5,709
vut	VUT	5,718	317	317	6,352
wbi	WBI	4,955	275	275	5,505
weh	WEH	5,995	333	333	6,661
wib	WIB	4,721	262	262	5,245
wja	WJA	1,145	63	63	1,271
wmw	WMW	10,657	592	592	11,841
wob	WOB	5,704	316	316	6,336
wof	WOF	1,570	87	87	1,744
xan	XAN	2,527	140	140	2,807
xed	XED	16,909	939	939	18,787
xmd	XMD	6,527	362	362	7,251
xmg	XMG	5,991	332	332	6,655
xnj	XNJ	7,482	415	415	8,312
xnz	XNZ	2,207	122	122	2,451
xog	XOG	18,419	1,023	1,023	20,465
xon	XON	22,339	1,241	1,241	24,821
xrb	XRB	5,743	318	318	6,379
xsm	XSM	22,341	1,241	1,241	24,823
xsn	XSN	1,535	85	85	1,705
xtc	XTC	6,204	344	344	6,892
xuo	XUO	6,014	334	334	6,682
yaf	YAF	5,308	294	294	5,896
yal	YAL	20,615	1,145	1,145	22,905
yam	YAM	5,433	301	301	6,035
yas	YAS	6,079	337	337	6,753
yat	YAT	5,219	289	289	5,797
yav	YAV	4,797	266	266	5,329
yba	YBA	6,356	353	353	7,062
yns	YNS	5,986	332	332	6,650
yom	YOM	11,084	615	615	12,314
yre	YRE	6,019	334	334	6,687
zaj	ZAJ	10,667	592	592	11,851
zak	ZAK	5,444	302	302	6,048
zga	ZGA	4,389	243	243	4,875
zin	ZIN	5,662	314	314	6,290
ziw	ZIW	5,591	310	310	6,211
Total		4,165,056	231,177	231,177	4,627,410

---

How to Load Subsets

You can load any specific African language subset directly with Hugging Face datasets:

from datasets import load_dataset

# 1. Load Twi (Akan) subset
dataset_tw = load_dataset("AfriSpeech/gloss-mt-train-fr", "tw")
print(dataset_tw["train"][0])

# 2. Load Swahili subset
dataset_sw = load_dataset("AfriSpeech/gloss-mt-train-fr", "sw")
print(dataset_sw["train"][0])

# 3. Load Amharic subset (if available)
dataset_amh = load_dataset("AfriSpeech/gloss-mt-train-fr", "amh")
print(dataset_amh["train"][0])

Data Fields

• input: The source sentence with in-place target glossary annotations (e.g. The <kannifoɔ> made a covenant...).
• tgt: The clean target sentence (e.g. Na kannifoɔ no ne Nyankopɔn yɛɛ apam...).
• src_raw: Original clean source sentence before glossary injection.
• tgt_ref: Original target reference sentence.
• inserted_terms: List of glossary terms injected into the source text.
• src_lang: Source language code.
• tgt_lang: Target language code.

---

Citation

@article{owusu2026afrigloss,
  title={Afri-Gloss-MT: In-Place Glossary Constraints for African Machine Translation},
  author={Owusu, Mich-Seth},
  journal={arXiv preprint},
  year={2026}
}