Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -12,9 +12,27 @@ WinoGrande train and development sets recast as coreference resolution as descri
|
|
12 |
|
13 |
```python
|
14 |
{
|
15 |
-
|
16 |
-
|
17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
18 |
}
|
19 |
```
|
20 |
|
|
|
12 |
|
13 |
```python
|
14 |
{
|
15 |
+
"id": str, # example id
|
16 |
+
"text": str, # untokenized example text
|
17 |
+
"sentences": [
|
18 |
+
{
|
19 |
+
"id": int, # sentence index
|
20 |
+
"text": str, # untokenized sentence text
|
21 |
+
"speaker": None, # speaker
|
22 |
+
"tokens": [
|
23 |
+
{
|
24 |
+
# keys are conllu columns: id, text, lemma, upos, xpos, feats, head, deprel, deps, misc
|
25 |
+
},
|
26 |
+
...
|
27 |
+
]
|
28 |
+
},
|
29 |
+
...
|
30 |
+
],
|
31 |
+
"coref_chains": List[List[List[int]]], # list of clusters, each cluster is a list of mentions, each mention is a span represented as [sent, start, end] inclusive
|
32 |
+
"genre": "crowdsourced",
|
33 |
+
"meta_data": {
|
34 |
+
"comment": "syntax_annotations=stanza|tokenizer=stanza|detokenizer=nltk",
|
35 |
+
},
|
36 |
}
|
37 |
```
|
38 |
|