Datasets:
matteogabburo
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -50,130 +50,6 @@ configs:
|
|
50 |
path: "por-test.jsonl"
|
51 |
- split: test_sp
|
52 |
path: "spa-test.jsonl"
|
53 |
-
- split: validation_clean_en
|
54 |
-
path: "eng-dev_clean.jsonl"
|
55 |
-
- split: validation_clean_de
|
56 |
-
path: "deu-dev_clean.jsonl"
|
57 |
-
- split: validation_clean_fr
|
58 |
-
path: "fra-dev_clean.jsonl"
|
59 |
-
- split: validation_clean_it
|
60 |
-
path: "ita-dev_clean.jsonl"
|
61 |
-
- split: validation_clean_po
|
62 |
-
path: "por-dev_clean.jsonl"
|
63 |
-
- split: validation_clean_sp
|
64 |
-
path: "spa-dev_clean.jsonl"
|
65 |
-
- split: test_clean_en
|
66 |
-
path: "eng-test_clean.jsonl"
|
67 |
-
- split: test_clean_de
|
68 |
-
path: "deu-test_clean.jsonl"
|
69 |
-
- split: test_clean_fr
|
70 |
-
path: "fra-test_clean.jsonl"
|
71 |
-
- split: test_cleanit
|
72 |
-
path: "ita-test_clean.jsonl"
|
73 |
-
- split: test_clean_po
|
74 |
-
path: "por-test_clean.jsonl"
|
75 |
-
- split: test_clean_sp
|
76 |
-
path: "spa-test_clean.jsonl"
|
77 |
-
- split: validation_noneg_en
|
78 |
-
path: "eng-dev_no_allneg.jsonl"
|
79 |
-
- split: validation_noneg_de
|
80 |
-
path: "deu-dev_no_allneg.jsonl"
|
81 |
-
- split: validation_noneg_fr
|
82 |
-
path: "fra-dev_no_allneg.jsonl"
|
83 |
-
- split: validation_noneg_it
|
84 |
-
path: "ita-dev_no_allneg.jsonl"
|
85 |
-
- split: validation_noneg_po
|
86 |
-
path: "por-dev_no_allneg.jsonl"
|
87 |
-
- split: validation_noneg_sp
|
88 |
-
path: "spa-dev_no_allneg.jsonl"
|
89 |
-
- split: test_noneg_en
|
90 |
-
path: "eng-test_no_allneg.jsonl"
|
91 |
-
- split: test_noneg_de
|
92 |
-
path: "deu-test_no_allneg.jsonl"
|
93 |
-
- split: test_noneg_fr
|
94 |
-
path: "fra-test_no_allneg.jsonl"
|
95 |
-
- split: test_noneg_it
|
96 |
-
path: "ita-test_no_allneg.jsonl"
|
97 |
-
- split: test_noneg_po
|
98 |
-
path: "por-test_no_allneg.jsonl"
|
99 |
-
- split: test_noneg_sp
|
100 |
-
path: "spa-test_no_allneg.jsonl"
|
101 |
-
- config_name: clean
|
102 |
-
data_files:
|
103 |
-
- split: train_en
|
104 |
-
path: "eng-train.jsonl"
|
105 |
-
- split: train_de
|
106 |
-
path: "deu-train.jsonl"
|
107 |
-
- split: train_fr
|
108 |
-
path: "fra-train.jsonl"
|
109 |
-
- split: train_it
|
110 |
-
path: "ita-train.jsonl"
|
111 |
-
- split: train_po
|
112 |
-
path: "por-train.jsonl"
|
113 |
-
- split: train_sp
|
114 |
-
path: "spa-train.jsonl"
|
115 |
-
- split: validation_clean_en
|
116 |
-
path: "eng-dev_clean.jsonl"
|
117 |
-
- split: validation_clean_de
|
118 |
-
path: "deu-dev_clean.jsonl"
|
119 |
-
- split: validation_clean_fr
|
120 |
-
path: "fra-dev_clean.jsonl"
|
121 |
-
- split: validation_clean_it
|
122 |
-
path: "ita-dev_clean.jsonl"
|
123 |
-
- split: validation_clean_po
|
124 |
-
path: "por-dev_clean.jsonl"
|
125 |
-
- split: validation_clean_sp
|
126 |
-
path: "spa-dev_clean.jsonl"
|
127 |
-
- split: test_clean_en
|
128 |
-
path: "eng-test_clean.jsonl"
|
129 |
-
- split: test_clean_de
|
130 |
-
path: "deu-test_clean.jsonl"
|
131 |
-
- split: test_clean_fr
|
132 |
-
path: "fra-test_clean.jsonl"
|
133 |
-
- split: test_cleanit
|
134 |
-
path: "ita-test_clean.jsonl"
|
135 |
-
- split: test_clean_po
|
136 |
-
path: "por-test_clean.jsonl"
|
137 |
-
- split: test_clean_sp
|
138 |
-
path: "spa-test_clean.jsonl"
|
139 |
-
- config_name: noneg
|
140 |
-
data_files:
|
141 |
-
- split: train_en
|
142 |
-
path: "eng-train.jsonl"
|
143 |
-
- split: train_de
|
144 |
-
path: "deu-train.jsonl"
|
145 |
-
- split: train_fr
|
146 |
-
path: "fra-train.jsonl"
|
147 |
-
- split: train_it
|
148 |
-
path: "ita-train.jsonl"
|
149 |
-
- split: train_po
|
150 |
-
path: "por-train.jsonl"
|
151 |
-
- split: train_sp
|
152 |
-
path: "spa-train.jsonl"
|
153 |
-
- split: validation_noneg_en
|
154 |
-
path: "eng-dev_no_allneg.jsonl"
|
155 |
-
- split: validation_noneg_de
|
156 |
-
path: "deu-dev_no_allneg.jsonl"
|
157 |
-
- split: validation_noneg_fr
|
158 |
-
path: "fra-dev_no_allneg.jsonl"
|
159 |
-
- split: validation_noneg_it
|
160 |
-
path: "ita-dev_no_allneg.jsonl"
|
161 |
-
- split: validation_noneg_po
|
162 |
-
path: "por-dev_no_allneg.jsonl"
|
163 |
-
- split: validation_noneg_sp
|
164 |
-
path: "spa-dev_no_allneg.jsonl"
|
165 |
-
- split: test_noneg_en
|
166 |
-
path: "eng-test_no_allneg.jsonl"
|
167 |
-
- split: test_noneg_de
|
168 |
-
path: "deu-test_no_allneg.jsonl"
|
169 |
-
- split: test_noneg_fr
|
170 |
-
path: "fra-test_no_allneg.jsonl"
|
171 |
-
- split: test_noneg_it
|
172 |
-
path: "ita-test_no_allneg.jsonl"
|
173 |
-
- split: test_noneg_po
|
174 |
-
path: "por-test_no_allneg.jsonl"
|
175 |
-
- split: test_noneg_sp
|
176 |
-
path: "spa-test_no_allneg.jsonl"
|
177 |
- config_name: en
|
178 |
data_files:
|
179 |
- split: train
|
@@ -222,102 +98,6 @@ configs:
|
|
222 |
path: "spa-dev.jsonl"
|
223 |
- split: test
|
224 |
path: "spa-test.jsonl"
|
225 |
-
- config_name: en_noneg
|
226 |
-
data_files:
|
227 |
-
- split: train
|
228 |
-
path: "eng-train.jsonl"
|
229 |
-
- split: validation
|
230 |
-
path: "eng-dev_no_allneg.jsonl"
|
231 |
-
- split: test
|
232 |
-
path: "eng-test_no_allneg.jsonl"
|
233 |
-
- config_name: de_noneg
|
234 |
-
data_files:
|
235 |
-
- split: train
|
236 |
-
path: "deu-train.jsonl"
|
237 |
-
- split: validation
|
238 |
-
path: "deu-dev_no_allneg.jsonl"
|
239 |
-
- split: test
|
240 |
-
path: "deu-test_no_allneg.jsonl"
|
241 |
-
- config_name: fr_noneg
|
242 |
-
data_files:
|
243 |
-
- split: train
|
244 |
-
path: "fra-train.jsonl"
|
245 |
-
- split: validation
|
246 |
-
path: "fra-dev_no_allneg.jsonl"
|
247 |
-
- split: test
|
248 |
-
path: "fra-test_no_allneg.jsonl"
|
249 |
-
- config_name: it_noneg
|
250 |
-
data_files:
|
251 |
-
- split: train
|
252 |
-
path: "ita-train.jsonl"
|
253 |
-
- split: validation
|
254 |
-
path: "ita-dev_no_allneg.jsonl"
|
255 |
-
- split: test
|
256 |
-
path: "ita-test_no_allneg.jsonl"
|
257 |
-
- config_name: po_noneg
|
258 |
-
data_files:
|
259 |
-
- split: train
|
260 |
-
path: "por-train.jsonl"
|
261 |
-
- split: validation
|
262 |
-
path: "por-dev_no_allneg.jsonl"
|
263 |
-
- split: test
|
264 |
-
path: "por-test_no_allneg.jsonl"
|
265 |
-
- config_name: sp_noneg
|
266 |
-
data_files:
|
267 |
-
- split: train
|
268 |
-
path: "spa-train.jsonl"
|
269 |
-
- split: validation
|
270 |
-
path: "spa-dev_no_allneg.jsonl"
|
271 |
-
- split: test
|
272 |
-
path: "spa-test_no_allneg.jsonl"
|
273 |
-
- config_name: en_clean
|
274 |
-
data_files:
|
275 |
-
- split: train
|
276 |
-
path: "eng-train.jsonl"
|
277 |
-
- split: validation
|
278 |
-
path: "eng-dev_clean.jsonl"
|
279 |
-
- split: test
|
280 |
-
path: "eng-test_clean.jsonl"
|
281 |
-
- config_name: de_clean
|
282 |
-
data_files:
|
283 |
-
- split: train
|
284 |
-
path: "deu-train.jsonl"
|
285 |
-
- split: validation
|
286 |
-
path: "deu-dev_clean.jsonl"
|
287 |
-
- split: test
|
288 |
-
path: "deu-test_clean.jsonl"
|
289 |
-
- config_name: fr_clean
|
290 |
-
data_files:
|
291 |
-
- split: train
|
292 |
-
path: "fra-train.jsonl"
|
293 |
-
- split: validation
|
294 |
-
path: "fra-dev_clean.jsonl"
|
295 |
-
- split: test
|
296 |
-
path: "fra-test_clean.jsonl"
|
297 |
-
- config_name: it_clean
|
298 |
-
data_files:
|
299 |
-
- split: train
|
300 |
-
path: "ita-train.jsonl"
|
301 |
-
- split: validation
|
302 |
-
path: "ita-dev_clean.jsonl"
|
303 |
-
- split: test
|
304 |
-
path: "ita-test_clean.jsonl"
|
305 |
-
- config_name: po_clean
|
306 |
-
data_files:
|
307 |
-
- split: train
|
308 |
-
path: "por-train.jsonl"
|
309 |
-
- split: validation
|
310 |
-
path: "por-dev_clean.jsonl"
|
311 |
-
- split: test
|
312 |
-
path: "por-test_clean.jsonl"
|
313 |
-
- config_name: sp_clean
|
314 |
-
data_files:
|
315 |
-
- split: train
|
316 |
-
path: "spa-train.jsonl"
|
317 |
-
- split: validation
|
318 |
-
path: "spa-dev_clean.jsonl"
|
319 |
-
- split: test
|
320 |
-
path: "spa-test_clean.jsonl"
|
321 |
---
|
322 |
|
323 |
## Dataset Description
|
@@ -334,13 +114,8 @@ For each language (English, French, German, Italian, Portuguese, and Spanish), w
|
|
334 |
- **validation** split
|
335 |
- **test** split
|
336 |
|
337 |
-
In addition, the validation and the test splits are available also in the following preprocessed versions:
|
338 |
-
|
339 |
-
- **noneg**: without questions with only negative answer candidates
|
340 |
-
- **clean**: without questions with only negative and only positive answer candidates
|
341 |
-
|
342 |
### How to load them:
|
343 |
-
To use these splits, you can use the following snippet of code replacing ``[LANG]`` with a language identifier (en, fr, de, it, po, sp)
|
344 |
|
345 |
```
|
346 |
from datasets import load_dataset
|
@@ -348,12 +123,6 @@ from datasets import load_dataset
|
|
348 |
# if you want the whole corpora
|
349 |
corpora = load_dataset("matteogabburo/mTRECQA")
|
350 |
|
351 |
-
# if you want the clean test and test sets
|
352 |
-
corpora = load_dataset("matteogabburo/mTRECQA", "clean")
|
353 |
-
|
354 |
-
# if you want the "no all negatives" validation and test sets
|
355 |
-
corpora = load_dataset("matteogabburo/mTRECQA", "noneg")
|
356 |
-
|
357 |
"""
|
358 |
if you want the default splits of a specific language, replace [LANG] with an identifier in: en, fr, de, it, po, sp
|
359 |
dataset = load_dataset("matteogabburo/mTRECQA", "[LANG]")
|
@@ -361,14 +130,6 @@ dataset = load_dataset("matteogabburo/mTRECQA", "[LANG]")
|
|
361 |
# example:
|
362 |
italian_dataset = load_dataset("matteogabburo/mTRECQA", "it")
|
363 |
|
364 |
-
|
365 |
-
"""
|
366 |
-
if you want the processed splits ("clean" and "no all negatives" sets), replace [LANG] with a language identifier and [VERSION] with "noneg" or "clean"
|
367 |
-
dataset = load_dataset("matteogabburo/mTRECQA", "[LANG]_[VERSION]")
|
368 |
-
"""
|
369 |
-
# example:
|
370 |
-
italian_clean_dataset = load_dataset("matteogabburo/mTRECQA", "it_clean")
|
371 |
-
|
372 |
```
|
373 |
|
374 |
|
|
|
50 |
path: "por-test.jsonl"
|
51 |
- split: test_sp
|
52 |
path: "spa-test.jsonl"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
- config_name: en
|
54 |
data_files:
|
55 |
- split: train
|
|
|
98 |
path: "spa-dev.jsonl"
|
99 |
- split: test
|
100 |
path: "spa-test.jsonl"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
101 |
---
|
102 |
|
103 |
## Dataset Description
|
|
|
114 |
- **validation** split
|
115 |
- **test** split
|
116 |
|
|
|
|
|
|
|
|
|
|
|
117 |
### How to load them:
|
118 |
+
To use these splits, you can use the following snippet of code replacing ``[LANG]`` with a language identifier (en, fr, de, it, po, sp)
|
119 |
|
120 |
```
|
121 |
from datasets import load_dataset
|
|
|
123 |
# if you want the whole corpora
|
124 |
corpora = load_dataset("matteogabburo/mTRECQA")
|
125 |
|
|
|
|
|
|
|
|
|
|
|
|
|
126 |
"""
|
127 |
if you want the default splits of a specific language, replace [LANG] with an identifier in: en, fr, de, it, po, sp
|
128 |
dataset = load_dataset("matteogabburo/mTRECQA", "[LANG]")
|
|
|
130 |
# example:
|
131 |
italian_dataset = load_dataset("matteogabburo/mTRECQA", "it")
|
132 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
133 |
```
|
134 |
|
135 |
|