References:
ady
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ady')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1666 |
- Features:
ang
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ang')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1867 |
- Features:
ara
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ara')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
4134 |
- Features:
arn
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/arn')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
26 |
- Features:
ast
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ast')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
436 |
- Features:
aze
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/aze')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
340 |
- Features:
bak
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/bak')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1084 |
- Features:
bel
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/bel')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1027 |
- Features:
ben
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ben')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
136 |
- Features:
bod
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/bod')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1335 |
- Features:
bre
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/bre')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
44 |
- Features:
bul
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/bul')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
2468 |
- Features:
cat
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/cat')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1547 |
- Features:
ces
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ces')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
5125 |
- Features:
chu
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/chu')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
152 |
- Features:
ckb
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ckb')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
274 |
- Features:
cor
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/cor')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
9 |
- Features:
crh
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/crh')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1230 |
- Features:
csb
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/csb')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
37 |
- Features:
cym
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/cym')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
183 |
- Features:
dan
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/dan')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
3193 |
- Features:
deu
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/deu')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
15060 |
- Features:
dsb
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/dsb')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
994 |
- Features:
ell
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ell')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
11906 |
- Features:
eng
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/eng')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
22765 |
- Features:
est
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/est')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
886 |
- Features:
eus
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/eus')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
26 |
- Features:
fao
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fao')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
3077 |
- Features:
fas
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fas')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
273 |
- Features:
fin
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fin')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'1' |
46152 |
'2' |
11491 |
- Features:
fra
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fra')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
7535 |
- Features:
frm
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/frm')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
603 |
- Features:
fro
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fro')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1700 |
- Features:
frr
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/frr')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
51 |
- Features:
fry
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fry')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
85 |
- Features:
fur
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/fur')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
168 |
- Features:
gal
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/gal')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
486 |
- Features:
gla
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/gla')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
73 |
- Features:
gle
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/gle')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
7464 |
- Features:
glv
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/glv')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1 |
- Features:
gmh
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/gmh')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
29 |
- Features:
gml
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/gml')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
52 |
- Features:
got
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/got')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
0 |
- Features:
grc
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/grc')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
2431 |
- Features:
hai
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/hai')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
41 |
- Features:
hbs
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/hbs')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
24419 |
- Features:
heb
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/heb')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
510 |
- Features:
hin
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/hin')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
258 |
- Features:
hun
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/hun')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
14892 |
- Features:
hye
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/hye')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
7033 |
- Features:
isl
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/isl')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
4775 |
- Features:
ita
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/ita')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
10009 |
- Features:
izh
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/izh')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
50 |
- Features:
kal
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kal')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
23 |
- Features:
kan
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kan')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
159 |
- Features:
kat
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kat')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
3782 |
- Features:
kaz
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kaz')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
26 |
- Features:
kbd
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kbd')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
250 |
- Features:
kjh
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kjh')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
75 |
- Features:
klr
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/klr')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
591 |
- Features:
kmr
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/kmr')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
15083 |
- Features:
krl
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/krl')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
20 |
- Features:
lat
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/lat')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
17214 |
- Features:
lav
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/lav')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
7548 |
- Features:
lit
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/lit')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
1458 |
- Features:
liv
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:universal_morphologies/liv')
- Description:
The Universal Morphology (UniMorph) project is a collaborative effort to improve how NLP handles complex morphology in the world’s languages.
The goal of UniMorph is to annotate morphological data in a universal schema that allows an inflected word from any language to be defined by its lexical meaning,
typically carried by the lemma, and by a rendering of its inflectional form in terms of a bundle of morphological features from our schema.
The specification of the schema is described in Sylak-Glassman (2016).
- License: CC-BY-SA-3.0
- Version: 1.0.0
- Splits:
Split | Examples |
---|---|
'train' |
203 |
- Features:
lld
Use the following command to load this dataset in TFD