Safetensors
gemma3_text

All data is taken from Fineweb-2 which consists of web-crawls (cutoff April 2024). The exact data can be obtained by running the code in the repository that will be released later.

Sizes of training data in number of UTF-8 characters:

Language Size
abt 6608062
aeb 49814258
amp 3737645
ape 4980833
arb 3242372164
azj 3242372164
bak 486333599
bbc 35774794
bdq 41451
bew 3242372164
bmr 2439163
bss 1731272
bzj 4036823
caf 909293
cak 14765286
cmn 3242372164
cnh 141157675
crn 5912477
crs 88583543
ctd 28218128
efi 17408332
ell 3242372164
eng 3242372164
ewe 35627506
gug 38458467
hau 3242372164
hil 123755905
hin 3242372164
ibo 333471679
ind 3242372164
kmr 3242372164
kua 14827393
kum 20332449
lat 3242372164
lus 230520812
mps 6855963
mri 534077417
mzn 2835443
nrm 3242372164
ory 3242372164
pan 3242372164
pol 3242372164
pon 12919138
qug 15887852
qxh 2452330
rus 3242372164
sat 4507904
sin 3242372164
sme 174411164
smo 417316517
sna 359593003
swh 3242372164
tur 3242372164
ubu 7941636
uig 603363013
vie 3242372164
wal 11207859
wol 18292867
wrk 2419613
yon 2356518
zai 6148374
total 68719476736

contact:

Downloads last month
4
Safetensors
Model size
1.0B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Dataset used to train 3CPO-models/nietsvermoedend-oneCluster-1b