Duplicates in the LAMA dataset

I observed duplicates in the LAMA probing dataset, see a minimal code below.

```
>>> import datasets
>>> dataset = datasets.load_dataset('lama')
No config specified, defaulting to: lama/trex
Reusing dataset lama (/home/anam/.cache/huggingface/datasets/lama/trex/1.1.0/97deffae13eca0a18e77dfb3960bb31741e973586f5c1fe1ec0d6b5eece7bddc)
>>> train_dataset = dataset['train']
>>> train_dataset[0]
{'description': 'language or languages a person has learned from early childhood', 'label': 'native language', 'masked_sentence': 'Louis Jules Trochu ([lwi ʒyl tʁɔʃy]; 12 March 1815 – 7 October 1896) was a [MASK] military leader and politician.', 'obj_label': 'French', 'obj_surface': 'French', 'obj_uri': 'Q150', 'predicate_id': 'P103', 'sub_label': 'Louis Jules Trochu', 'sub_surface': 'Louis Jules Trochu', 'sub_uri': 'Q441235', 'template': 'The native language of [X] is [Y] .', 'template_negated': '[X] is not owned by [Y] .', 'type': 'N-1', 'uuid': '40b2ed1c-0961-482e-844e-32596b6117c8'}
>>> train_dataset[1]
{'description': 'language or languages a person has learned from early childhood', 'label': 'native language', 'masked_sentence': 'Louis Jules Trochu ([lwi ʒyl tʁɔʃy]; 12 March 1815 – 7 October 1896) was a [MASK] military leader and politician.', 'obj_label': 'French', 'obj_surface': 'French', 'obj_uri': 'Q150', 'predicate_id': 'P103', 'sub_label': 'Louis Jules Trochu', 'sub_surface': 'Louis Jules Trochu', 'sub_uri': 'Q441235', 'template': 'The native language of [X] is [Y] .', 'template_negated': '[X] is not owned by [Y] .', 'type': 'N-1', 'uuid': '40b2ed1c-0961-482e-844e-32596b6117c8'}
```

I checked the original data available at https://dl.fbaipublicfiles.com/LAMA/data.zip. This particular duplicated comes from:
```
{"uuid": "40b2ed1c-0961-482e-844e-32596b6117c8", "obj_uri": "Q150", "obj_label": "French", "sub_uri": "Q441235", "sub_label": "Louis Jules Trochu", "predicate_id": "P103", "evidences": [{"sub_surface": "Louis Jules Trochu", "obj_surface": "French", "masked_sentence": "Louis Jules Trochu ([lwi \u0292yl t\u0281\u0254\u0283y]; 12 March 1815 \u2013 7 October 1896) was a [MASK] military leader and politician."}, {"sub_surface": "Louis Jules Trochu", "obj_surface": "French", "masked_sentence": "Louis Jules Trochu ([lwi \u0292yl t\u0281\u0254\u0283y]; 12 March 1815 \u2013 7 October 1896) was a [MASK] military leader and politician."}]}
```

What is the best way to deal with these duplicates if I want to use `datasets` to probe with LAMA?